Whitepaper: A Comparative Study of PCoIP and Blast Protocols for Horizon View

1. Historical Background

Evolution of Remote Display: Early remote display technologies like Microsoft RDP (1990s) and Citrix ICA laid the groundwork for VDI (Virtual Desktop Infrastructure) by transmitting graphical interfaces over networks. Teradici’s PC-over-IP (PCoIP) emerged in the mid-2000s as a breakthrough in delivering a high-performance, pixel-perfect remote desktop experience (Teradici – Wikipedia) (Microsoft Word – TER0806005 Issue 1 – WP-Network Infrastructure Requirements_FINAL_July4_bhz.doc). PCoIP was distinctive for offloading display rendering to a dedicated protocol, initially in hardware, to send only pixels (not application data) over IP networks. VMware quickly recognized its potential – in 2008 VMware licensed PCoIP for integration into VMware View (later Horizon) (Teradici – Wikipedia), delivering a richer user experience than prior methods. By 2009, VMware View 4.0 shipped with PCoIP as a primary display protocol, marking a major milestone in VDI history (A complete history of VMware and VDI).

PCoIP Timeline: Teradici (founded 2004) debuted its first PCoIP products in 2007 – a host card for servers and a “portal” zero client device with a custom Teradici TERA chip implementing the PCoIP protocol (Teradici – Wikipedia). These enabled high-performance remote graphics over standard networks. In 2013, Amazon Web Services adopted Teradici’s PCoIP for Amazon WorkSpaces cloud desktops, underscoring industry trust in PCoIP’s quality (Teradici – Wikipedia). Over the next decade, Teradici evolved PCoIP from purely hardware-based to also software-based (PCoIP Soft Client and agent), extending it to cloud and workstation remoting use cases. Notably, Teradici introduced PCoIP Ultra enhancements around 2019 to support 4K/UHD workloads and GPU encoding (using NVIDIA NVENC) for modern demands (PCoIP Ultra Technical Series Part 1: Top-Level Architecture) (PCoIP Ultra Technical Series Part 1: Top-Level Architecture). In 2021, HP Inc. acquired Teradici (Teradici – Wikipedia), integrating PCoIP into its HP Anyware remote computing portfolio and continuing protocol development. PCoIP’s evolution reflects major advancements in remote display tech: from LAN-only hardware solutions to efficient software codecs for WAN and cloud, all while maintaining a focus on high fidelity and security.

VMware Blast Timeline: VMware’s own protocol, originally called Blast, emerged amid the shift to mobile and cloud. VMware first introduced Blast in 2013 as an HTML5 browser-based remote console option (Announcing end of support for PCoIP in VMware Horizon – VMware End-User Computing Blog) (Understand the basics of VMware Blast Extreme vs. PCoIP | TechTarget). This early “Blast” allowed using just a web browser to access virtual desktops, but had limitations. VMware then developed Blast Extreme, a fully featured display protocol, released with VMware Horizon 7 in 2016 (In detail — VMware’s Blast Extreme protocol) (Announcing end of support for PCoIP in VMware Horizon – VMware End-User Computing Blog). Blast Extreme was built from scratch to leverage modern codecs (initially H.264) and adapt to a wide range of devices and networks. Over time, VMware rapidly improved Blast – adding support for UDP transport, adaptive networking (BEAT – Blast Extreme Adaptive Transport), newer codecs like H.265/HEVC, and achieving feature parity with PCoIP. By the late 2010s, Blast became the default protocol in Horizon. In 2023, VMware (now under Broadcom) announced PCoIP support will be officially dropped from Horizon after 2025, signaling full commitment to Blast going forward (Announcing end of support for PCoIP in VMware Horizon – VMware End-User Computing Blog) (Announcing end of support for PCoIP in VMware Horizon – VMware End-User Computing Blog). This transition underscores how Blast Extreme matured to meet enterprise expectations that PCoIP once set.

Key Milestones: Below is a summary timeline of major milestones for PCoIP and Blast:

• 2004–2007 (Teradici/PCoIP Launch): Teradici founded (2004) and unveils PCoIP hardware solution in 2007 (host card + zero client) (Teradici – Wikipedia), pioneering host rendering and pixel streaming.
• 2008–2009 (VMware Adopts PCoIP): VMware licenses PCoIP in 2008 and integrates it into VMware View 4.0 (2009), bringing high-performance remote graphics to mainstream VDI (Teradici – Wikipedia).
• 2013 (VMware Blast Introduced): VMware releases the first Blast protocol (HTML5-based) and later that year Amazon launches WorkSpaces service using PCoIP (Teradici – Wikipedia), highlighting PCoIP’s cloud relevance.
• 2015–2016 (Blast Extreme & Advancements): VMware develops Blast Extreme (Horizon 7 in 2016) as a mobile- and cloud-optimized protocol built on H.264, while Teradici refines PCoIP in software. Industry competition (Citrix HDX, etc.) also pushes innovation.
• 2019 (PCoIP Ultra): Teradici’s PCoIP Ultra enhancements (beta in 2019) add multi-codec and GPU encode support for 4K video, reducing CPU use and bandwidth for demanding workloads (PCoIP Ultra Technical Series Part 1: Top-Level Architecture) (PCoIP Ultra Technical Series Part 1: Top-Level Architecture).
• 2021 (HP Acquisition): HP Inc. acquires Teradici (Oct 2021) (Teradici – Wikipedia), ensuring continued investment in PCoIP (now part of HP Anyware) for hybrid work and multi-cloud deployments.
• 2023 (Broadcom/VMware & EOL Announcement): Broadcom’s acquisition of VMware looms, and VMware announces end-of-support for PCoIP in Horizon by end of 2025, citing Blast’s evolution as the preferred protocol (Announcing end of support for PCoIP in VMware Horizon – VMware End-User Computing Blog). HP Anyware and AWS continue to offer PCoIP outside of VMware’s ecosystem.
• Present Day: PCoIP and Blast Extreme represent two generations of remote display tech – one born in bespoke silicon for on-prem VDI, the other in software for the mobile-cloud era. Together, their history traces the industry’s journey toward high-performance, secure remote computing across any network.

2. Technical Deep Dive

PCoIP Protocol Architecture and Technology

Architecture Overview: PCoIP is a host-rendering, pixel-streaming protocol designed to transmit an entire desktop experience over IP with fidelity rivaling a local PC. In a typical PCoIP deployment, the host system (virtual desktop or workstation) runs a PCoIP agent or has a PCoIP host card, which captures the display output, encodes it, and sends it as an encrypted pixel stream. On the client side, a PCoIP client (software client or Teradici zero client hardware) receives and decodes the pixels to present the remote desktop. All processing of the OS and graphics happens on the host – the client is essentially a thin decoder. This architecture means no application data or graphics primitives traverse the network – only rendered pixels, USB signals, and audio streams are transmitted (About PCoIP Technology – HP Anyware Architecture Guide) (About PCoIP Technology – HP Anyware Architecture Guide). Original Teradici implementations achieved this via dedicated silicon: the PCoIP Host Processor (in the server) would intercept the DVI/DisplayPort output and USB I/O, compress them in real time, and send to a PCoIP Portal Processor in the zero client, which decrypts and displays video and routes peripheral data back to the host (Microsoft Word – TER0806005 Issue 1 – WP-Network Infrastructure Requirements_FINAL_July4_bhz.doc) (Microsoft Word – TER0806005 Issue 1 – WP-Network Infrastructure Requirements_FINAL_July4_bhz.doc). Modern PCoIP software agents perform the same functions in software with CPU/GPU acceleration as available.

Encoding and Compression: PCoIP uses an adaptive, multi-codec compression engine purpose-built for computer imagery (text, UI, 3D graphics) rather than a single video codec. It analyzes content and dynamically adjusts codecs, quality levels, and frame rates to balance clarity and bandwidth. For example, PCoIP can transmit text and fine details in a truly lossless manner when needed – it supports a “build-to-lossless” mode where the image is progressively refined until it matches the original pixel-for-pixel (About PCoIP Technology – HP Anyware Architecture Guide). This is critical for use cases like medical imaging or CAD, where even minor compression artifacts are unacceptable. PCoIP’s codec arsenal includes specialized algorithms for different image regions: e.g. detecting static text or UI elements and compressing them losslessly, while using optimized lossy compression for natural images or video. It does not rely on H.264 for desktop graphics; instead, it uses advanced mix of run-length encoding, build-to-lossless bitmaps, and when motion is present, efficient codecs that can update changes rapidly. Teradici’s design continuously monitors network conditions and “adaptively controls image data quality and update rate” in real time (Microsoft Word – TER0806005 Issue 1 – WP-Network Infrastructure Requirements_FINAL_July4_bhz.doc). If bandwidth is constrained or latency rises, PCoIP will temporarily reduce image quality (e.g. compress more aggressively or lower frame rate) to maintain interactivity, then improve quality (even to lossless) when conditions allow (Microsoft Word – TER0806005 Issue 1 – WP-Network Infrastructure Requirements_FINAL_July4_bhz.doc). This results in very efficient bandwidth use – idle or static screens use almost no bandwidth except for perfection refinements, while fast motion is compressed on the fly. Later enhancements in PCoIP Ultra introduced the option to use H.264/HEVC for full-screen video content or leverage CPU AVX2 instructions and GPU encoding, essentially adding an H.264 codec to PCoIP’s toolkit for scenarios where a video codec is advantageous (PCoIP Ultra Technical Series Part 1: Top-Level Architecture) (PCoIP Ultra Technical Series Part 1: Top-Level Architecture). This multi-codec approach is a key differentiator: “Blast Extreme uses a single H.264 video codec to compress the entire display, while the PCoIP protocol leverages multiple codecs optimized for text and graphics” (Blast Extreme Requires GPUs to Deliver a Competitive User Experience). Thus, PCoIP can deliver sharp text and lossless images alongside efficient video playback without user tuning – the protocol intelligently chooses the best compression method per content type.

Transport and Network Efficiency: PCoIP is inherently a UDP-based protocol, engineered for real-time performance (About PCoIP Technology – HP Anyware Architecture Guide). It operates over UDP (port 4172 by default) for the bulk pixel stream, with an AES-encrypted session. Control signaling (for session handshake, etc.) can run over TCP or a secure tunnel, but all interactive media (display, audio, etc.) uses UDP to minimize latency (About PCoIP Technology – HP Anyware Architecture Guide) (Data protection in Amazon WorkSpaces – Amazon WorkSpaces). UDP allows PCoIP to avoid the head-of-line blocking and throughput constraints of TCP, instead implementing its own loss recovery and congestion management optimized for interactive traffic. PCoIP monitors packet loss and available capacity and will adjust its output accordingly. For example, if network bandwidth drops, PCoIP dynamically reduces image quality and frame updates to fit the pipe, rather than stalling (Microsoft Word – TER0806005 Issue 1 – WP-Network Infrastructure Requirements_FINAL_July4_bhz.doc). It “adapts its traffic profile to fluctuations in available network resources” so that even in congested networks with many users, it fairly shares bandwidth and maintains responsiveness (Microsoft Word – TER0806005 Issue 1 – WP-Network Infrastructure Requirements_FINAL_July4_bhz.doc). This adaptive behavior was quite advanced for its time, ensuring smooth performance from LAN to WAN. PCoIP’s default MTU and packet sizing are tuned to balance latency and efficiency (one analysis noted PCoIP tends to send consistently sized UDP packets ~1200 bytes, buffering updates until it can send an optimal payload to maximize throughput without fragmentation (Remote protocols benchmarking, Citrix, VMware and RDP–Part One PCoIP vs Blast Extreme | Marius Sandbu – IT blog)). In practice, PCoIP can deliver a full desktop with surprisingly low bandwidth for typical office work – Teradici often touted that PCoIP can provide a high-quality experience for office productivity in ~150–250 kbps in steady state, scaling up for video or rich media as needed (peak usage in the order of a few Mbps for 1080p video). PCoIP also supports advanced features like QoS tagging and client-side shaping to further manage network impact. The result is a protocol that is very bandwidth-efficient across a range of scenarios. In fact, when Blast (H.264) is not GPU-accelerated, tests show it often consumes significantly more bandwidth than PCoIP for equivalent tasks (Blast Extreme Requires GPUs to Deliver a Competitive User Experience) (Blast Extreme Requires GPUs to Deliver a Competitive User Experience) (since H.264 doesn’t natively economize on text and screen changes as well as PCoIP’s tailored codecs).

Performance Characteristics: PCoIP was designed for low latency and high frame rates to make remote desktops “perception-free.” By using UDP and very tight encoding loops (the original hardware could compress and send a frame in under a single display refresh interval (Microsoft Word – EVGA PCoIP_User_Guide.doc)), PCoIP minimizes lag. In optimal conditions (LAN or low-latency WAN), PCoIP can deliver 60 fps HD video and interactive 3D graphics that feel local. It supports multiple monitors (up to four at high resolution) and synchronizes outputs for a seamless experience. PCoIP zero clients are famed for their smoothness – because the decoding is done in hardware with fixed function ASICs, frame latency is minimal and consistent. Interactive latency (mouse movements, etc.) is very low; Teradici literature notes the total round-trip can be just a few milliseconds plus network ping, meaning users barely perceive they’re not on a local machine (Microsoft Word – EVGA PCoIP_User_Guide.doc) (Microsoft Word – EVGA PCoIP_User_Guide.doc). Even in software implementations, PCoIP is highly optimized with SIMD instructions. Where PCoIP historically lagged is in pure video playback efficiency – a complex scene of full-screen video at 60 fps could tax CPU resources if no hardware assist, leading to frame rate drops. The Ultra enhancements addressed this by offloading video to H.264 on GPU when available (PCoIP Ultra Technical Series Part 1: Top-Level Architecture). With PCoIP Ultra, performance for high-motion content improved dramatically (4K video at 60 fps became feasible with low overhead). Benchmarks from HP show that PCoIP with CPU encoding consumes less server CPU and bandwidth than Blast/H.264 software encoding for typical 2D workloads (Blast Extreme Requires GPUs to Deliver a Competitive User Experience) (Blast Extreme Requires GPUs to Deliver a Competitive User Experience), translating to higher user density per server in VDI. Overall, PCoIP’s performance excels in delivering sharp imagery and consistent interaction for knowledge workers, and with recent upgrades it can also handle heavier media workloads well.

Security Features: Security is a core design aspect of PCoIP. The protocol was built to ensure that no sensitive data ever leaves the data center, only encrypted pixels. In the original architecture, the PCoIP host and client communicate over two secure channels: a control channel and a media channel. The control channel (for session establishment, USB device authorization, etc.) is secured with mutual certificate-based authentication, ensuring the client and host verify each other (Microsoft Word – EVGA PCoIP_User_Guide.doc). The media/data channel carrying the pixels, audio, etc. is protected with robust encryption (initially AES-128, later AES-256 in GCM mode) (Data protection in Amazon WorkSpaces – Amazon WorkSpaces). All traffic is encrypted in transit – Teradici obtained FIPS 140-2 certification for its PCoIP implementations, attesting to strong cryptography (Data protection in Amazon WorkSpaces – Amazon WorkSpaces) ([PDF] FIPS 140-2 Non-Proprietary Security Policy). PCoIP sessions can be configured to use AES-256-bit encryption for high security environments (the default in some systems is AES-128-GCM for performance, but policies can enforce 256-bit) (Data protection in Amazon WorkSpaces – Amazon WorkSpaces). Because PCoIP runs over UDP (which is not inherently secure), the encryption and authentication are built into the protocol (often using DTLS or similar under the hood when software-based). Another security advantage is the stateless client design: a Teradici zero client has no OS or persistent storage, eliminating risk of malware on the client side (Microsoft Word – EVGA PCoIP_User_Guide.doc) (Microsoft Word – EVGA PCoIP_User_Guide.doc). It simply decrypts and displays pixels. There are no attack surfaces like registries or local drives, and if a zero client is stolen, no data is on it. USB peripheral redirection is tightly controlled – admins can authorize or block specific USB device classes to prevent data leakage (e.g. disallow mass storage) (Microsoft Word – EVGA PCoIP_User_Guide.doc USB devices)) (Microsoft Word – EVGA PCoIP_User_Guide.doc). PCoIP also integrates with smart cards and CAC for user authentication, often used in government deployments. In summary, PCoIP provides enterprise-grade security: “The PCoIP Portal is a stateless device with no local data… Host/Portal communication is secured using digital certificates for mutual authentication, and the media data stream is encrypted by an AES algorithm” (Microsoft Word – EVGA PCoIP_User_Guide.doc) (Microsoft Word – EVGA PCoIP_User_Guide.doc). Many organizations have used PCoIP without a VPN, leveraging its native encryption and the fact that it can be constrained to specific network ports (4172 UDP/TCP) through firewalls (Data protection in Amazon WorkSpaces – Amazon WorkSpaces). This simplifies remote access while still protecting data. To date, there have been no known major security breaches of the PCoIP protocol – its closed, purpose-built nature and Teradici’s ongoing updates (often in coordination with VMware and now HP) have maintained a strong security record.

VMware Blast Extreme Protocol Architecture and Technology

Architecture Overview: VMware Blast Extreme (often just “Blast”) is a display remoting protocol introduced to replace or augment PCoIP in Horizon deployments. Unlike PCoIP’s proprietary roots, Blast was designed using more standard codecs and transport to leverage broadly available hardware acceleration. In a Horizon environment, the Blast Extreme protocol is implemented by the Horizon Agent on the virtual desktop or RDSH server, and the Horizon Client on the end-user device. The protocol can operate through VMware’s Connection Server and Unified Access Gateway (for brokered, secure access) or directly between agent and client on the LAN. Blast’s architecture is modular: it consists of an encoder component on the server, which captures the graphical output of the desktop and encodes it (primarily using video codecs), and a decoder on the client. One of VMware’s design goals was unification – Blast had to support not just thick clients but also thin clients, mobile devices, and even web browsers. Thus, the protocol was built to use common network ports and standard encryption (SSL/TLS), easing traversal of firewalls and compatibility with existing infrastructure (VMware Horizon 7 Blast Extreme Primer—Everything an Admin Needs to Know – VMware End-User Computing Blog) (In detail — VMware’s Blast Extreme protocol). A Blast session by default runs over TCP port 8443 or 443 (SSL), and can optionally (or adaptively) use UDP 8443 for the bulk traffic, encapsulated in DTLS for security (In detail — VMware’s Blast Extreme protocol) (Blast Extreme Polices | The SLOG – SimonLong/Blog). In essence, VMware layered Blast on top of widely-used web protocols: it can share port 443 with HTTPS traffic (via a Blast Secure Gateway) to simplify network requirements (VMware Horizon 7 Blast Extreme Primer—Everything an Admin Needs to Know – VMware End-User Computing Blog). This architecture choice made it straightforward for clients – any device that can make an HTTPS (TLS) connection and decode H.264 video can potentially use Blast. Indeed, Blast is supported on Windows, macOS, Linux, iOS, Android, ChromeOS, and even HTML5 browsers (VMware Horizon 7 Blast Extreme Primer—Everything an Admin Needs to Know – VMware End-User Computing Blog) (In detail — VMware’s Blast Extreme protocol). Internally, Blast comprises several codec options and channels (for instance, separate channels for imaging, audio, USB redirection, etc., similar to RDP’s virtual channels). By Horizon 7, VMware ensured Blast could provide all the virtual desktop features (multiple monitors, audio in/out, USB, smart cards, printing, etc.) that users expected from PCoIP (In detail — VMware’s Blast Extreme protocol). This parity was achieved by building on VMware’s virtualization stack (for example, leveraging VMware virtual audio and USB redirection components while simply transporting their data within Blast’s stream). From an architecture perspective, Blast Extreme is software-based – it does not require special hardware on clients, though it takes advantage of hardware (GPUs) if present.

Codec and Compression Methods: The cornerstone of Blast Extreme is the use of the H.264 video codec (also known as AVC) for encoding the screen graphics (VMware Horizon 7 Blast Extreme Primer—Everything an Admin Needs to Know – VMware End-User Computing Blog) (In detail — VMware’s Blast Extreme protocol). H.264 was chosen because it is ubiquitous and highly efficient for a wide range of imagery. Modern CPUs and GPUs have dedicated H.264 decoding capabilities, especially in mobile devices – this meant a Blast client could offload video decoding to hardware, drastically reducing CPU usage on phones, tablets, and thin clients (In detail — VMware’s Blast Extreme protocol) (VMware Horizon 7 Blast Extreme Primer—Everything an Admin Needs to Know – VMware End-User Computing Blog). On the server side, if an NVIDIA GRID GPU is present, Blast can offload encoding to the GPU’s NVENC engine, saving server CPU (In detail — VMware’s Blast Extreme protocol). This approach aligns with trends in the mid-2010s: video codecs had become the de facto standard for streaming not just media but also real-time game streams, etc., so VMware leveraged that momentum. Blast’s default mode is to capture the entire desktop as a series of video frames and encode them in H.264. It is essentially treating the desktop like a video feed. However, VMware recognized that one codec might not fit all scenarios, so Blast also has additional encoding modes: JPG/PNG for static images, and a “Blast Codec” (proprietary adaptive codec) introduced later. In fact, the Blast protocol can dynamically switch between codecs based on content or administrator policy. For example, Blast can use a JPEG/PNG codec for lossless still image quality when the screen is mostly static (this is inherited from the earlier Blast HTML5 client) (In detail — VMware’s Blast Extreme protocol%2C Linux)). VMware documentation describes an “encoder switch” that can “dynamically switch between the Blast Codec, JPG/PNG, and H.264 codec” depending on content type (40 Tips for Optimizing VMware Horizon VDI’s Blast Extreme Protocol). The Blast Codec mentioned is a VMware-proprietary codec optimized for text and UI (likely a lightweight run-length or similar codec to provide clearer text than raw H.264). By default, though, Blast uses lossy H.264 compression for most content (In detail — VMware’s Blast Extreme protocol). Administrators can configure Blast to be lossless, but that typically incurs higher bandwidth usage (In detail — VMware’s Blast Extreme protocol). With Horizon 7.3 and above, VMware also added optional H.265/HEVC support (when using clients and GPUs that support it) for even better compression of high-res visuals. In summary, Blast’s compression methodology is video-centric: it strives to send delta frames at a high frequency (targeting 30–60 FPS), using inter-frame compression to reduce redundant data. This works extremely well for photorealistic and continuous motion (like video playback), often outperforming PCoIP in those areas when GPU-accelerated (VMware Horizon Blast Extreme Acceleration with NVIDIA GRID | VMware End-User Computing Blog) (VMware Horizon Blast Extreme Acceleration with NVIDIA GRID | VMware End-User Computing Blog). One trade-off is that H.264 is inherently lossy for text and may introduce slight blurriness in fine lines unless in lossless mode (Understand the basics of VMware Blast Extreme vs. PCoIP | TechTarget) (). VMware mitigated this by allowing fine-tuning (the Blast Extreme settings include quality levels and the option to constrain quantization for sharper text).

Transport and Networking: Blast Extreme is versatile in transport – it supports TCP, UDP, and adaptive schemes to deal with varying network conditions. Initially, Blast was often configured to use TCP for ease of traversal (everything over TCP 443), but VMware later enabled Blast Extreme Adaptive Transport (BEAT) by default (in Horizon 7.1+), which attempts UDP if possible and falls back to TCP if not (Deep Dive into VMware Horizon Blast Extreme Adaptive Transport) (Blast Extreme Polices | The SLOG – SimonLong/Blog). Under BEAT, Blast will prefer UDP 8443 (or UDP 22443 internally) with a DTLS encrypted channel for low-latency data, while still using TCP 443/8443 for control and virtual channel traffic (What is Blast Extreme Network Intelligent Transport (BENIT)?) (VMware Horizon 7 Blast Extreme Primer—Everything an Admin Needs to Know – VMware End-User Computing Blog). This hybrid approach gives “best of both” – the reliability of TCP for things like printing or file transfer, and the low-latency benefits of UDP for screen updates. Blast also has congestion control and error recovery mechanisms. H.264 itself can tolerate some packet loss (resulting in momentary artifacts), and Blast will drop frames if network is constrained rather than queue them, to keep latency low. One notable difference is that Blast, when using TCP, relies on TCP flow control which can increase latency on high-loss networks; PCoIP never does that as it’s UDP-only. However, VMware’s guidance is to use UDP whenever possible for Blast Extreme to avoid TCP drawbacks (Deep Dive into VMware Blast Extreme Network Intelligent Transport). With UDP, Blast can handle up to ~25% packet loss by degrading quality before the user experience breaks down (In detail — VMware’s Blast Extreme protocol – Basvankaam.com). Blast’s network ports are firewall-friendly: by tunneling through port 443 (HTTPS) via the Unified Access Gateway, a Blast session can be made to look like ordinary TLS traffic, simplifying remote access setup (VMware Horizon 7 Blast Extreme Primer—Everything an Admin Needs to Know – VMware End-User Computing Blog). Security-wise, Blast uses TLS/DTLS with modern ciphers (AES-128/AES-256) so that its transport is secure and often compliant with enterprise encryption standards. By design, Blast does not require additional VPN if used with VMware’s Access Point/UAG – it was meant to be internet-facing in Horizon Cloud services, etc., with encryption in place and even NSA Suite B cryptography available in some client versions. Overall, Blast’s adaptability in networking made it suitable for both LAN (where it can achieve high FPS, even 4K resolution at modest bandwidth with H.264) and WAN (where it will throttle frame rate or use a “thinwire” mode if needed).

Performance Characteristics: VMware engineered Blast Extreme to be competitive with or superior to PCoIP in performance, especially in newer use cases. In practice, Blast’s performance can be split into two scenarios: with GPU offload and without. With GPU offload (NVENC) on the server, Blast can encode frames very quickly and with less CPU, achieving high frame rates for 3D and video. Tests showed that using NVIDIA NVENC, Blast Extreme reduced the total latency by ~50 ms compared to PCoIP, and by 27 ms compared to Blast using CPU encoding (VMware Horizon Blast Extreme Acceleration with NVIDIA GRID | VMware End-User Computing Blog). The reduction comes from faster encoding and perhaps better pipelining. In the same tests (ESRI ArcGIS workload), Blast with GPU also boosted frame rates by ~37% over CPU encoding (VMware Horizon Blast Extreme Acceleration with NVIDIA GRID | VMware End-User Computing Blog). These results indicate that in GPU-accelerated environments, Blast delivers excellent performance for graphically intensive apps. It can comfortably support 1080p/60 or even 4K/30 for CAD, video editing, etc., on a single vGPU. Without GPU offload, Blast relies on CPU to encode H.264, which can be intensive. In such cases, PCoIP’s efficient algorithms may have an edge in not maxing out CPU. Indeed, one study found that with no GPU, Blast’s H.264 encoding consumed about 20% more CPU than PCoIP for the same workload, potentially reducing the number of VMs per host by a similar 20% (Blast Extreme Requires GPUs to Deliver a Competitive User Experience). Blast was also observed to use more bandwidth in pure CPU mode – roughly 2× the bandwidth of PCoIP for typical office tasks when both use software encoding (Blast Extreme Requires GPUs to Deliver a Competitive User Experience). This is attributed to H.264 not being as efficient on sharp text and repetitive patterns, causing it to send more data for the same visual quality (Blast Extreme Requires GPUs to Deliver a Competitive User Experience). However, when Blast can use a GPU, it flips – the protocol tends to send a higher quality (less compressed) stream since the GPU can handle it, which ironically can increase bandwidth usage (one test showed Blast with NVENC used 5× more bandwidth than PCoIP with hardware assist in certain cases (Blast Extreme Requires GPUs to Deliver a Competitive User Experience) (Blast Extreme Requires GPUs to Deliver a Competitive User Experience)). VMware’s own benchmarking, though, demonstrated that in a controlled test with 19 virtual desktops running a GIS workload, Blast with NVENC used ~19% less bandwidth than PCoIP at comparable quality (VMware Horizon Blast Extreme Acceleration with NVIDIA GRID | VMware End-User Computing Blog). This suggests that performance can vary by content: for video or continuously changing graphics, Blast (especially HEVC) can be more bandwidth-efficient, whereas for mostly static or office work, PCoIP’s adaptive codecs can use less bandwidth. Ultimately, both protocols perform well within typical enterprise network constraints. Latency-wise, Blast on UDP is very close to PCoIP’s responsiveness. Users on modern Horizon clients often report minimal observable difference in interactivity between Blast and PCoIP, especially on LAN. On high-latency links (e.g. 200+ ms), both will suffer, but Blast’s ability to fall back to TCP ensures a session stays connected (with increased lag) whereas PCoIP might struggle or disconnect in extremely lossy conditions (since it doesn’t have built-in forward error correction beyond re-transmits). Blast supports up to 4 monitors at 4K resolution (with Horizon 8 and later), matching PCoIP’s multi-monitor support. For audio and realtime communications, VMware developed features like Blast Extreme Adaptive Transport and Network Intelligence (BENIT) to improve stability for VoIP and webcam redirection over Blast (Horizon Blast Extreme UDP with BEAT Support Functionality in BIG …). As a result, VMware was able to optimize unified communications (Skype, Teams) on Blast to a point where it’s now the preferred protocol for those use cases. One area where PCoIP long had an advantage is ultra-low latency in zero client scenarios – a PCoIP zero client can sometimes feel more “crisp” than a software client due to the minimal decode latency and USB responsiveness. Blast does not have a bespoke zero client (though some thin clients now embed Linux that runs a Horizon client), so there is a slight increase in client-side latency (in the order of a few milliseconds). Teradici highlighted that “Blast clients deliver noticeably higher interactive latency than PCoIP Zero Clients” in some comparisons (Blast Extreme Does Not Measure up to PCoIP – Teradici). In practice, this difference is small and often outweighed by Blast’s other benefits, but in environments like stock trading or simulation, a tuned PCoIP zero client might still beat Blast by a few milliseconds in reaction time. Summing up, Blast Extreme’s performance is highly tunable and has improved with each Horizon release. By leveraging standard codecs and hardware acceleration, it achieved parity with PCoIP for most tasks by around 2018. Today, Blast can handle anything from typing in Word (with clear text after refinement) to streaming 3D CAD or full-motion video.

(VMware Horizon Blast Extreme Acceleration with NVIDIA GRID | VMware End-User Computing Blog) Blast Extreme vs. PCoIP bandwidth utilization: An example performance comparison from VMware shows host bandwidth over time for a graphic-intensive workload. In this test, Horizon with Blast Extreme (GPU-accelerated, green line) was able to transmit the workload with slightly lower cumulative bandwidth than Horizon with PCoIP (blue line). Offloading Blast encoding to GPU reduced latency by ~51 ms and increased frame rates, while also cutting bandwidth ~19% compared to PCoIP in this scenario (VMware Horizon Blast Extreme Acceleration with NVIDIA GRID | VMware End-User Computing Blog). (Lower is better on the graph — note the green line peaks earlier and drops off, indicating efficient burst encoding.)

Security Features: Blast Extreme, as part of Horizon, inherits VMware’s security architecture. All Blast sessions are encrypted in transit, typically using TLS 1.2 with strong ciphers (AES-128 by default, and AES-256 if configured). When Blast uses TCP 443, it is essentially running over HTTPS – the Horizon Connection Server and Unified Access Gateway can terminate or proxy these TLS connections just like a web server (VMware Horizon 7 Blast Extreme Primer—Everything an Admin Needs to Know – VMware End-User Computing Blog). For UDP mode, Blast uses DTLS (Datagram TLS) on UDP 8443/22443 to ensure encryption and integrity of the stream. Authentication for Blast is handled by Horizon: users authenticate to the Connection Server (via Active Directory credentials, smart card, SAML, etc.), and then a Blast session is initiated. Blast doesn’t have a separate mutual auth mechanism like PCoIP’s host cards did; instead, it relies on the broker to only allow authorized clients to connect (the Blast session is tied to a session token). From a data perspective, Blast only transmits display and peripheral data, similar to PCoIP. One key difference is that Blast sessions can be brokered through VMware’s gateway which provides an additional layer of security (DMZ component). This means Blast can be deployed with no inbound ports opened directly to desktops – a Unified Access Gateway in the DMZ handles all external Blast traffic over TLS and then forwards it internally. Such architecture was possible with PCoIP only via special security servers since PCoIP wasn’t natively web-friendly. In terms of encryption strength, both protocols meet strict security requirements (for example, Horizon clients and agents (with Blast) have been certified in Common Criteria evaluations ([PDF] VMware Horizon Client 8 2209 (Horizon 8.7) Security Target – NIAP)). Neither Blast nor PCoIP have known vulnerabilities in the cryptographic design publicly reported. That said, security-conscious customers sometimes prefer PCoIP because of the zero-client option (no attack surface on client). With Blast, the client is often a general-purpose OS which must be secured (though locked-down thin OS or mobile devices can be hardened adequately). Blast’s reliance on TLS means it’s as secure as HTTPS – which is generally considered very secure when using modern cipher suites. VMware continues to update Blast’s security; for example, if a vulnerability is found in OpenSSL or the TLS stack, a Horizon client/agent update is issued. Both protocols support features like USB device filtering, and both can be configured to comply with FIPS cryptography standards (Horizon can be set to use FIPS mode ciphers for Blast). In summary, security is a draw between the two protocols: Blast uses industry-standard TLS/DTLS encryption and integrates with VMware’s authentication and access control, whereas PCoIP uses proprietary AES encryption and often runs on isolated hardware – but both achieve a high level of security for government and enterprise use. One consideration is that after VMware drops PCoIP, Blast will receive all the ongoing security enhancements and patches in Horizon, whereas PCoIP’s future updates will come from HP for non-Horizon deployments.

3. Vendor Landscape

Teradici (HP) and the PCoIP Ecosystem: PCoIP was created by Teradici, a Canadian company that drove its adoption through partnerships. Teradici’s technology became embedded in many vendors’ offerings. VMware was an early adopter (licensing PCoIP for VMware Horizon), but other major tech players also integrated PCoIP: Dell (Wyse) and HP sold PCoIP zero clients and even integrated Teradici chips into all-in-one monitors (Teradici – Wikipedia). OEMs like Samsung, Fujitsu, and Amulet Hotkey included Teradici hardware in their endpoint devices or blade workstations (Teradici – Wikipedia). Later, Amazon AWS became a significant PCoIP user – Amazon WorkSpaces has long used PCoIP as the primary protocol for delivering cloud desktops to users (Teradici – Wikipedia). This broad adoption indicates a robust ecosystem: PCoIP was not tied to a single vendor’s stack, which appealed to customers who wanted multi-vendor flexibility. After HP’s acquisition of Teradici in 2021, PCoIP development and support continue under HP Anyware (the new name for Teradici’s Cloud Access Software) (Announcing end of support for PCoIP in VMware Horizon – VMware End-User Computing Blog). HP Anyware positions PCoIP as a universal remote workstation solution that works across public clouds, on-prem data centers, and hybrid environments – distinguishing it from VMware’s Blast which is specific to Horizon. In HP’s vision, an organization might use PCoIP to remote high-end workstations (for engineering or media production) from any cloud to any device, with HP providing the software (agent and client) and licensing. They emphasize that “unlike Blast Extreme, PCoIP is supported on all clouds: public, private or hybrid” (PCoIP… The Freedom To Choose – Teradici – HP Anyware). The vendor landscape for PCoIP thus includes HP as the protocol owner and software provider, AWS as a cloud service provider utilizing PCoIP (under license), and a host of client device manufacturers (10ZiG, ClearCube, Dell, IGEL, etc.) who ensure their hardware/firmware supports PCoIP. Even post-HP acquisition, firmware updates for Teradici-based zero clients are being delivered (now labeled as HP Anyware clients). This indicates continued investment in the PCoIP ecosystem for at least the near term.

VMware (Broadcom) and Blast: VMware developed Blast Extreme primarily for its own Horizon VDI product, and that remains the only major platform where Blast is used. VMware’s aim was to control its destiny in display protocol technology (eliminating dependency on Teradici) and optimize for emerging needs like mobile clients and cloud desktops. Over the years, VMware has improved Blast in tandem with Horizon releases – adding features, improving performance, and working closely with GPU vendors (NVIDIA, AMD, Intel) to enhance Blast. Notably, VMware partnered with NVIDIA to utilize GPUs for Blast, which was highlighted when NVIDIA GRID vGPU came to market (In detail — VMware’s Blast Extreme protocol). As of Horizon 8 (and Horizon Cloud on Azure), Blast is the default and preferred protocol. VMware has also cultivated a thin client ecosystem around Blast: many thin client vendors (Dell Wyse, HP, Stratodesk, IGEL) provide Horizon Clients that support Blast Extreme out of the box. Some newer zero clients even support Blast Extreme by running a lightweight Linux-based client internally – for example, Dell and ClearCube released firmware for their Teradici-based zero clients to also connect via Blast (taking advantage that those devices have H.264 decode capability) (Zero Client – Stateless, Secure, VMware Ready PCoIP TERA2). This cross-compatibility eased the transition for customers with large PCoIP zero client deployments – they could start using Blast without replacing all endpoints, in some cases.

With Broadcom’s acquisition of VMware (pending or completed in 2023–2024), VMware’s EUC division is expected to continue focusing on enterprise customers and high-performance use cases. Broadcom has already indicated no intent to support PCoIP long-term (evidenced by the Horizon announcement of PCoIP EoL) (Announcing end of support for PCoIP in VMware Horizon – VMware End-User Computing Blog). Instead, VMware (Broadcom) will push Blast Extreme as the sole protocol in future Horizon versions. This likely means even stronger vendor optimization and consolidation around Blast. For instance, future enhancements might include better integration with Broadcom’s hardware (if any relevant technology exists) or continued collaboration with GPU makers to refine Blast.

Other Vendors and Alternatives: While Teradici/HP and VMware are the primary players for PCoIP and Blast respectively, it’s worth noting the broader competitive landscape. Citrix, VMware’s longtime rival in VDI, has its HDX/ICA protocol, which also uses multiple codecs and adaptive display technology. Many comparisons include Citrix HDX vs PCoIP vs Blast – each has its pros/cons, but Citrix remains a separate ecosystem (not using PCoIP or Blast). Microsoft has enhanced RDP over the years (with RemoteFX and now AVC/H.264 modes) and even introduced a new protocol for Windows 365 cloud PCs. These alternatives mean customers have choices beyond PCoIP/Blast if they choose different platforms. However, organizations invested in VMware Horizon or Teradici’s solutions are typically deciding between PCoIP and Blast. Some niche vendors have built on one of these protocols: e.g. Leostream (a connection broker) can broker sessions using either PCoIP or Blast by leveraging the respective clients. Amulet Hotkey, which provides remote workstation solutions, historically used PCoIP in its hardware (and recently has explored alternative protocols as VMware shifts strategies) (Why You Need To Consider Alternative Protocols To PCoIP). Another example is AWS, which apart from PCoIP also developed a proprietary WorkSpaces Streaming Protocol (WSP) in recent years – indicating even cloud providers sometimes seek custom protocols for better control. Nonetheless, Teradici’s PCoIP still powers tens of thousands of AWS WorkSpaces sessions daily, and HP Anyware is being adopted in industries like media/entertainment for remote editing and in government for secure remote access. The market impact of these protocols can be seen in how they enabled new use cases: PCoIP enabled “zero client” deployments in high-security environments (no Windows endpoints at desks, only PCoIP clients), improving security and manageability. Blast Extreme, on the other hand, has helped VMware extend Horizon to mobile and BYOD users, contributing to VDI being viable for a remote workforce. Adoption trends show that since 2016, Blast Extreme usage has grown rapidly and now dominates new Horizon deployments, while PCoIP usage has stagnated or declined in the VMware context. According to VMware, the vast majority of Horizon customers have transitioned to Blast as of the Horizon 8 era, which justified the decision to phase out PCoIP (Announcing end of support for PCoIP in VMware Horizon – VMware End-User Computing Blog) (Announcing end of support for PCoIP in VMware Horizon – VMware End-User Computing Blog). Meanwhile, PCoIP finds new life via HP, targeting non-VDI use cases (like remote physical workstations, Linux desktops, etc., which VMware’s solution doesn’t target as much). We also see convergence in client devices: many endpoint vendors advertise compatibility with “VMware Blast and PCoIP” together, ensuring their thin clients can connect to either protocol (e.g. IGEL OS endpoints, 10ZiG devices, and ClearCube clients all support both). This indicates that the industry doesn’t view it as an either-or at the endpoint level – one device can handle both protocols, giving customers flexibility.

In terms of market impact, VMware’s introduction of Blast forced Teradici to innovate (hence PCoIP Ultra) and also drove down some costs (as VMware no longer needed to pay Teradici licensing fees for new Horizon seats). Teradici pivoting to HP and the cloud shows how the technology is being repositioned. The upcoming end-of-2025 Horizon PCoIP support sunset is a critical point: after that, any VMware/Broadcom Horizon customer must use Blast (or RDP) only. We can expect HP to woo some of those customers toward HP Anyware if they truly need PCoIP for certain use cases, but many will simply stick with the VMware Blast trajectory.

Broadcom’s Role: The prompt mentions VMware/Broadcom together – Broadcom’s acquisition means that Blast Extreme is effectively now a Broadcom-owned technology (via VMware). Broadcom has a history of focusing on profitability and may not have specific expertise in remote display, but by consolidating on one protocol (Blast), it simplifies development and support. Broadcom also owns CA Technologies, which has a product called CA ASM (Secure Remote Access) – but there’s no indication of overlap with Blast. The more interesting Broadcom angle is whether they will integrate Blast with other Broadcom enterprise offerings (network optimizations, for example). It’s speculative, but Broadcom could potentially optimize Blast for certain network hardware or include it in a broader remote work solution.

Summary of Vendor Stance: HP Teradici champions PCoIP as a cloud-agnostic, ultra-secure, high-fidelity solution (especially for “power users” in graphics-intensive fields). VMware/Broadcom champion Blast Extreme as the simplified, all-purpose protocol for virtual apps and desktops in enterprises, with broad client support and ongoing innovation. Other vendors like ClearCube, Dell, IGEL, 10ZiG remain somewhat neutral, ensuring their endpoints support both protocols to cater to customer demand. Amazon continues to license and use PCoIP (with no indication of switching to Blast – AWS would presumably use their own WSP or NICE DCV if not PCoIP). In the DaaS (Desktop-as-a-Service) market, VMware’s Horizon Cloud uses Blast, AWS uses PCoIP, Citrix Cloud uses HDX – so the competition remains fierce but fragmented. Importantly, customer adoption tends to follow the vendor default: many VMware Horizon customers used PCoIP from 2009–2016 because it was the default; today, most use Blast because it is now the default and recommended setting (Announcing end of support for PCoIP in VMware Horizon – VMware End-User Computing Blog). As such, we see Blast’s share growing significantly in the installed base of virtual desktop protocols, while Teradici’s PCoIP finds a niche in specialized or non-VMware deployments.

4. Comparison Analysis – PCoIP vs. Blast Extreme

To provide a direct comparison, this section evaluates PCoIP and Blast across multiple dimensions:

• Performance & Latency: Both protocols are highly optimized for low latency, but their performance can diverge depending on scenario. In low-motion, office application use, PCoIP often delivers equal or better user experience with very crisp text and minimal bandwidth. Its progressive build approach means even on a slower link, text and UI become sharp after a brief moment. Blast Extreme in lossy mode might show minor compression on text unless tuned for lossless. However, Blast was shown to achieve lower end-to-end latency than PCoIP when GPU encoding is used – in one test, Blast reduced overall latency by ~51 ms versus PCoIP (VMware Horizon Blast Extreme Acceleration with NVIDIA GRID | VMware End-User Computing Blog). This suggests that for high-motion or 3D use cases with GPU assist, Blast can surpass PCoIP in responsiveness (encoding frames faster). For software-only scenarios, PCoIP’s single-image (intraframe) compression can be faster to encode than CPU-based H.264, potentially giving PCoIP a slight edge in interactive latency when no hardware assist is present (Blast Extreme Requires GPUs to Deliver a Competitive User Experience) (Blast Extreme Requires GPUs to Deliver a Competitive User Experience). In terms of frame rate, both can deliver 60 fps video if bandwidth and server resources allow. PCoIP adapts frame rate based on network (it might dip to e.g. 30 fps on constrained links), whereas Blast will try to maintain frame rate and drop quality first. Users have reported that Blast feels “snappier” on mobile devices due to hardware decoding (freeing the CPU) (Understand the basics of VMware Blast Extreme vs. PCoIP | TechTarget), whereas on a zero client, PCoIP feels extremely snappy (no OS overhead). In summary, with modern deployments (and especially if GPUs are available), Blast Extreme has caught up to PCoIP’s legendary performance, even excelling in some areas like smooth video playback and lower client-side CPU usage (Remote protocols benchmarking, Citrix, VMware and RDP–Part One PCoIP vs Blast Extreme | Marius Sandbu – IT blog). But PCoIP remains exceptionally interactive on LANs and proven in very demanding visualization tasks (film production, etc.), so performance is very close, with the tie-breaker often being hardware support (if you have Teradici hardware, PCoIP shines; if you have NVIDIA GPUs and H.264 endpoints, Blast shines).
• Bandwidth Efficiency: PCoIP is generally more bandwidth-efficient for typical knowledge worker workloads thanks to its adaptive codecs. It sends data only as needed to refine the display, and nothing when the screen is static. Blast’s default H.264 will continuously send video frames (even if highly similar, it still has some minimal frame size), which can lead to higher baseline bandwidth usage. Empirical comparisons show that “Blast H.264 uses twice the network bandwidth as PCoIP when both are being encoded in software” for an office productivity workload (Blast Extreme Requires GPUs to Deliver a Competitive User Experience). Teradici claims PCoIP can reduce bandwidth by up to 70% compared to Blast in some cases by using its advanced compression profiles (Teradici Blog | Blast Extreme Vs PCoIP – HP Anyware). However, when more motion or graphical complexity is involved (e.g. video playback or 3D rotation), Blast can become competitive or even more efficient – especially if using HEVC or if Blast dynamically lowers quality. In one VMware test with heavy GIS rendering, Blast (with H.264) used ~19% less bandwidth than PCoIP to achieve the same user experience (VMware Horizon Blast Extreme Acceleration with NVIDIA GRID | VMware End-User Computing Blog). It’s worth noting those tests involved many users and presumably PCoIP’s build-to-lossless sending more data to refine images. In general, PCoIP excels at bandwidth conservation for text, spreadsheets, email, etc., often using only a few tens of kilobits when little changes on screen (Microsoft Word – TER0806005 Issue 1 – WP-Network Infrastructure Requirements_FINAL_July4_bhz.doc). Blast excels at efficiently compressing full-motion video; its inter-frame compression can avoid repeatedly sending identical background pixels, for instance. With the advent of PCoIP Ultra (with H.264) and Blast HEVC, both protocols now leverage similar video compression for motion, so their video bandwidth usage can be comparable. But if bandwidth is a primary concern (like remote users on limited broadband or satellite links), PCoIP’s finely adaptive nature and ability to restrict to very low bandwidth while still updating screen content may be advantageous. Also, PCoIP has a unique ability to prioritize image fidelity when bandwidth allows – it will ensure the final drawn image is perfect given enough time, whereas Blast in lossy mode might never perfectly resolve pixel-by-pixel accuracy. That said, Blast can be configured with “Build-to-Lossless = On” in Horizon, which then makes it send lossless key frames (with a bandwidth cost). In summary, out-of-the-box PCoIP tends to consume less bandwidth for the same task in many everyday scenarios (Blast Extreme Requires GPUs to Deliver a Competitive User Experience), but Blast is not far behind and can even be more efficient in media-heavy uses or when GPU-accelerated (due to shorter bursts of high-quality frames). Administrators can tune both protocols via GPOs – for example, limit max bandwidth, turn on/off image refinement – to meet bandwidth constraints. Both protocols perform well on WAN, but PCoIP was historically known to function acceptably even in sub-1 Mbps environments (by heavily compressing and sacrificing frame rate if needed).
• Latency and Network Resilience: Latency tolerance is crucial for remote desktops. PCoIP was designed for LAN and WAN use and can work over high-latency links, but very high latency (e.g. >150 ms) can start to impact user experience as it has no mechanism to hide the fundamental round-trip delay on interactive actions. Blast, using TCP, has the drawback that high latency will directly add to round-trip (and TCP may introduce additional delay due to acknowledgments). However, Blast’s adaptive transport (UDP) makes it behave similarly to PCoIP under the hood in high-latency scenarios – dropping frames rather than waiting. Both protocols will show degraded responsiveness if latency is extremely high (because you can’t cheat physics), but they try to mitigate it. Blast’s advantage is that with UDP it can even handle out-of-order packets and continue, whereas PCoIP might be more sensitive to consistent packet order (Teradici doesn’t publish details, but presumably they also handle out-of-order reasonably). In terms of network loss, PCoIP can handle some packet loss by design (with some quality degradation) and has congestion avoidance to prevent self-induced loss (Microsoft Word – TER0806005 Issue 1 – WP-Network Infrastructure Requirements_FINAL_July4_bhz.doc). Blast over UDP (with forward error correction off) will show artifacting if packets drop, but it tolerates a decent amount (as noted, up to 25% loss with UDP using H.264 before things really break down) (In detail — VMware’s Blast Extreme protocol – Basvankaam.com). If Blast is forced to TCP on a lossy link, performance can suffer (TCP will retransmit and pause the stream, causing stutters). PCoIP doesn’t retransmit dropped frames (it prefers to move on with newer data), which can result in a better user experience on a lossy, high-latency link because it won’t stall – you might see a momentary reduction in quality but interaction continues. Blast’s BEAT tries to emulate that behavior by quickly switching to UDP when it detects issues. Overall, both protocols are robust on suboptimal networks, but PCoIP’s UDP-only, stateless nature historically gave it a reputation for coping well with inconsistent networks (e.g. 4G cellular or long-distance links). VMware’s improvements to Blast (like BENIT – Blast Network Intelligent Transport) are closing that gap by dynamically adjusting transport. If latency is very low (LAN environment), both deliver excellent performance; if latency is moderate (50–100 ms, typical remote office to cloud datacenter), users likely wouldn’t notice a difference in interactivity between the two. At very high latency (250+ ms, e.g. intercontinental), neither is ideal for real-time work, but Blast might have an edge by using client-side caching for video frames to smooth playback, whereas PCoIP might become choppy (conservative frame dropping). In practice, network conditions are rarely the deciding factor between these protocols anymore – both can be optimized for WAN and both can leverage VPN or SD-WAN acceleration if needed. It’s more about how they degrade: PCoIP will degrade image quality (blur) but keep latency low; Blast might retain image clarity longer but then drop frames or pause if TCP is in play.
• Hardware/Client Compatibility: This is a key difference area. PCoIP supports a unique class of hardware clients – Teradici-based zero clients – which have no OS and purely decode PCoIP. These clients cannot run Blast Extreme (unless they have been updated with some firmware that can run a Horizon Linux client, as a few models have done). If an organization has many PCoIP Zero Clients (e.g. in trading floors or healthcare), they can only use PCoIP to connect natively. (Some newer zero clients have added Blast support, but those are effectively not “zero” in the traditional sense; they have an embedded system that can handle H.264). Blast requires a software client – either Horizon Client on Windows/Linux/macOS, or mobile app, or HTML5. Most thin clients today (Linux-based or Windows IoT) can support both PCoIP and Blast by installing both Teradici and VMware client components. Compatibility-wise, PCoIP had an advantage in the past for specialized scenarios: for example, PCoIP could be integrated into secure KVM switches and IP phone devices (there were phones with built-in PCoIP client to give a combined voice+VDI terminal). Blast being newer doesn’t have such integrations widely. On the other hand, Blast works on a wider range of consumer devices – any modern device with a web browser can access Blast (via Horizon’s HTML Access), which wasn’t possible with PCoIP (PCoIP had no web client). Also, older devices that can’t run Teradici’s client but have a generic RDP client can sometimes still use RDP as a fallback in Horizon, but not PCoIP – so Blast has essentially replaced RDP as that “universally accessible” option, since H.264 decode is ubiquitous. In terms of server hardware, PCoIP originally required no GPU – it could use CPU to compress images – but if you wanted to accelerate heavy 3D, Teradici offered a PCoIP Hardware Accelerator (APEX card) and now relies on CPU AVX2 or optional NVENC. Blast practically requires GPU for heavy 3D if you want best results (otherwise CPU gets quite loaded). So if a deployment has no GPUs at all, running many users with Blast might mean higher CPU usage per VM, whereas PCoIP might be slightly lighter per frame on CPU (except in video scenarios). USB and peripheral support is equivalent for both in VMware Horizon context (since the Horizon agent’s USB redirection works regardless of protocol). Audio is supported bidirectionally in both; PCoIP can deliver up to 7.1 surround, Blast can also deliver HD audio. Both support multiple monitors: up to 4 monitors at 4K with Blast (Horizon 8) and PCoIP supports at least 4 (Teradici’s latest agents support 4 at 4K as well). One small difference: PCoIP’s host cards had direct support for analog video (early ones) and some unusual peripherals in certain industries (like serial port redirection) which Blast might need third-party solutions for. But for most compatibility questions, if you’re in the Horizon ecosystem, both protocols are supported on all the same client platforms (Windows, Mac, Linux, iOS, Android, ChromeOS thin clients). Outside Horizon, PCoIP is available via HP Anyware clients (Windows, Linux, Mac) – which are also fairly broad, but Blast is not available outside Horizon. Integration with brokers: PCoIP can work with non-VMware brokers (Leostream, AWS WorkSpaces, etc.), whereas Blast is tightly integrated with Horizon’s broker. This means if a company wanted a mix of different VDI solutions, PCoIP could be the common protocol across them (to a degree), but Blast cannot be used with anything except VMware Horizon.
• Security Considerations: Both protocols offer strong security but with different philosophies. PCoIP’s security approach is to minimize attack surface: using locked-down clients (or carefully audited software client) and keeping all data in the data center. It has proven secure enough for high-security government use (Teradici’s firmware got FIPS 140-2 validation ([PDF] FIPS 140-2 Non-Proprietary Security Policy) and is used in classified environments where data separation is critical). PCoIP Zero Clients can enforce smart card auth natively and are often certified for DoD use. Blast’s security approach leverages the robust security of TLS and modern encryption suites. Horizon environment can be configured with enterprise security features (e.g. integration with RSA SecurID, True SSO, etc.) that apply to Blast sessions. One security difference might be patching and provenance: PCoIP being proprietary meant fewer researchers looked at it, whereas Blast uses common components (OpenSSL, etc.) that are widely scrutinized. Either way, no known exploits specifically break Blast or PCoIP encryption. Vulnerabilities that have occurred tended to be in the surrounding infrastructure – e.g. a vulnerability in Horizon Connection Server or a weak cipher configuration. In 2020, for instance, organizations had to ensure TLS 1.2 was enforced for Blast to prevent any downgrade issues. PCoIP in VMware Horizon can also be impacted by such config (like default cipher was AES-128, which some wanted to up to AES-256). One advantage of Blast using TLS is easier inspection – organizations can use SSL inspection appliances (if they terminate and re-encrypt, though that’s rare for VDI traffic) or at least detect Blast traffic on 443. PCoIP’s UDP on 4172 is opaque and might be blocked by some security policies unless exceptions are made. So from a firewall perspective, Blast is easier to allow (looks like HTTPS) whereas PCoIP might require opening additional UDP ports. Regarding endpoint security, PCoIP Zero Clients are very secure endpoints by design (no malware risk locally), while Blast requires a general-purpose device which must be secured (anti-malware, OS updates, etc.). However, in corporate managed environments, thin clients running Blast can be stripped down and locked almost as tightly. Session security: Both protocols can enforce stringent encryption. By Horizon 7.10, VMware even supported NSA Suite B cryptography for Blast (elliptic curve certs, etc., primarily for government). Teradici likewise supports High Security Mode with AES-256-GCM only. We can conclude that both meet enterprise security requirements, with PCoIP perhaps having an edge for zero-trust endpoint scenarios, and Blast having an edge in ease of integration (using standard TLS, fewer custom ports).
• Use-Case Suitability: The choice between PCoIP and Blast often comes down to use-case and environment. PCoIP is well-suited for:
  • Ultra-secure and regulated industries: e.g. government, defense contractors, financial traders – places that deployed thousands of zero clients for security and manageability. They value the stateless client and the proven track record of PCoIP. Also, industries like media & entertainment or oil & gas, where color accuracy and lossless image quality are needed (e.g. post-production editing, geospatial satellite imaging) – PCoIP’s lossless capability ensures the remote image is pixel-perfect (About PCoIP Technology – HP Anyware Architecture Guide).
  • High-end workstations and dedicated remote graphics: Many studios use PCoIP to let artists access powerful GPUs in a data center from home via HP Anyware, since PCoIP handles dual 4K monitors and Wacom tablets etc. with ease. The “workstation remoting” use-case (1:1 users to machines) has been a PCoIP stronghold.
  • Low-bandwidth or high-loss links: remote sites with poor connectivity might get a better experience with PCoIP’s aggressive optimizations. PCoIP also can be easier to deploy in cloud-agnostic ways (HP Anyware can broker connections in multi-cloud), giving it an edge for hybrid cloud architectures where you don’t want to be tied to VMware’s stack.
  Blast Extreme is well-suited for:
  • General office VDI and large enterprise deployments: It’s optimized for tasks like Office apps, web browsing, video conferencing on virtual desktops, etc. With broad device support, employees can use Blast from their laptops, tablets, or repurposed PCs easily. The unified protocol reduces complexity for IT (one protocol to tune, one to troubleshoot).
  • Mobile and BYOD scenarios: Because Blast leverages native decoders, an iPad or Android phone can run a Horizon Blast session with minimal battery drain, giving near-native feeling. VMware explicitly optimized Blast for mobile networks and battery life (Understand the basics of VMware Blast Extreme vs. PCoIP | TechTarget) (Understand the basics of VMware Blast Extreme vs. PCoIP | TechTarget).
  • Cloud deployments and modern apps: If an organization is using VMware Horizon on Azure/AWS or Horizon DaaS, Blast is the supported protocol (PCoIP is not supported in Horizon Cloud Next-gen) (Announcing end of support for PCoIP in VMware Horizon – VMware End-User Computing Blog). Also, for published applications (RDSH apps via Horizon), Blast performs very well and simplifies access via browser or thin client.
  • Multimedia and Unified Communications: As mentioned, features like Microsoft Teams optimization have been added to Blast. VMware’s collaboration with Zoom and Teams to allow VDI offloading works over Blast. PCoIP can also support these (with Teradici’s collaboration with e.g. Cisco VXME in the past), but VMware is investing more in Blast for these use-cases now.

In essence, PCoIP often gets the nod in specialty use cases requiring absolute fidelity or specialized client hardware, whereas Blast is the default for mainstream VDI and cloud desktop use due to its flexibility and VMware’s backing.

5. Benefits and Drawbacks

PCoIP – Strengths and Weaknesses

Benefits/Strengths:

• Exceptional Image Quality (Lossless Support): PCoIP can gradually refine the display to a pixel-perfect replica of the host image (About PCoIP Technology – HP Anyware Architecture Guide). Text and graphics are crystal clear, with no compression artifacts when build-to-lossless is enabled. This is crucial for scenarios like medical imaging, print design, GIS, where fidelity is paramount. Competing H.264-based protocols struggle to offer true lossless delivery () ().
• Efficient Bandwidth Utilization: PCoIP intelligently adapts to network bandwidth. It sends highly compressed updates for changing parts of the screen and very little data when screens are static, making it very efficient in steady-state. Studies show it often uses significantly less bandwidth than Blast for typical office work (Blast Extreme Requires GPUs to Deliver a Competitive User Experience). It also has built-in congestion management to fairly share bandwidth in multi-user networks (Microsoft Word – TER0806005 Issue 1 – WP-Network Infrastructure Requirements_FINAL_July4_bhz.doc), preventing any one session from overwhelming the link.
• Low Latency, High Responsiveness: Designed to minimize delay, PCoIP over UDP yields fast response times. There’s no TCP ack overhead, and the protocol was tuned for LAN conditions – remote sessions “feel” near-local in many cases. PCoIP Zero Clients further reduce client-side latency (no OS, instantaneous decode), offering an excellent interactive experience for users (often cited as most “PC-like” experience by VDI admins) (Understand the basics of VMware Blast Extreme vs. PCoIP | TechTarget) (Understand the basics of VMware Blast Extreme vs. PCoIP | TechTarget). It handles input (keyboard/mouse) and output very smoothly, even under moderate network jitter.
• Strong Security & Stateless Endpoints: PCoIP endpoints can be completely stateless (no Windows or Linux OS). This dramatically reduces security risk at the endpoint – nothing to patch, no data exposure if a device is stolen. The protocol itself uses AES 256-bit encryption (with options for NSA Suite B ciphers) and has been FIPS 140-2 validated in implementations (Data protection in Amazon WorkSpaces – Amazon WorkSpaces) ([PDF] FIPS 140-2 Non-Proprietary Security Policy). All USB/peripheral traffic is carried within the encrypted session and can be filtered by policy, preventing unauthorized devices. PCoIP effectively keeps all sensitive data in the data center; only pixels leave, which is inherently safer against data exfiltration.
• Multi-Vendor Support & Flexibility: As a Teradici (now HP) offering, PCoIP is not tied to one VDI broker. It’s supported by VMware Horizon (until 2025) and by AWS WorkSpaces, and via HP Anyware it can be deployed in bespoke remote workstation setups. Many client devices from different vendors support PCoIP, giving customers choice. It’s cloud-agnostic – HP Anyware can enable PCoIP on AWS, Azure, GCP, or private cloud, which appeals to those avoiding lock-in (PCoIP… The Freedom To Choose – Teradici – HP Anyware). This flexibility is a strategic benefit if an organization wants the same protocol across multiple environments (e.g. on-prem VMware + cloud AWS) without being forced into one vendor’s stack.
• Proven Reliability at Scale: PCoIP has over a decade of field use in large deployments (tens of thousands of seats). It’s a mature protocol that has shown robust stability. Many organizations have built workflows around PCoIP (for example, in animation studios, dozens of artists collaborate remotely with PCoIP daily). Its longevity and refinements mean most “kinks” have been ironed out over time. Features like multi-monitor sync, USB device support, etc., are very well-implemented from years of feedback.
• Hardware Offload Options: For those who need it, PCoIP offers unique hardware offloads – e.g. the Teradici PCoIP Hardware Accelerator (APEX card) that could be added to servers to reduce CPU usage by handling encoding of changing pixels, or the use of Teradici PCoIP Remote Workstation Cards in physical workstations to offload all encoding. These can improve performance without requiring a full-blown GPU, and in some comparisons are more cost-effective per user than adding a GPU for Blast (Blast Extreme Requires GPUs to Deliver a Competitive User Experience) (Blast Extreme Requires GPUs to Deliver a Competitive User Experience). This scalability option (add an APEX card to boost PCoIP for many VMs) is a unique benefit in certain scenarios.

Drawbacks/Weaknesses:

• Dependence on Teradici/HP Ecosystem (Licensing): Outside of VMware Horizon or AWS (which bundle PCoIP licensing), using PCoIP requires licensing HP Anyware. This is an additional cost – HP Anyware is typically licensed per concurrent user. In contrast, Blast is included in the Horizon license. If an organization switches to a platform where PCoIP isn’t bundled, they’d have to purchase licenses to keep using it. With VMware planning to remove PCoIP, those who still want it may face extra procurement and integration steps to use HP’s solution alongside or instead of VMware’s.
• Higher CPU Requirements for Video/Graphical Workloads: While PCoIP is efficient for static and office tasks, it can become CPU-intensive when remoting high-framerate video or very graphically rich content entirely in software. Historically, playing 1080p video over PCoIP on a CPU-bound server could reduce frame rates and strain the VM, because PCoIP would be trying to encode lots of screen changes. Without GPU assist, PCoIP’s older algorithm might not compress motion as compactly as H.264. Teradici did address this with PCoIP Ultra (using H.264 for video), but that requires updated agents and is mostly targeted at standalone HP Anyware deployments (VMware’s Horizon PCoIP did not get the Ultra enhancements) (PCoIP Ultra Technical Series Part 1: Top-Level Architecture) (PCoIP Ultra Technical Series Part 1: Top-Level Architecture). So, in a Horizon setting, PCoIP encoding is single-threaded and was last optimized years ago (PCoIP Ultra Technical Series Part 1: Top-Level Architecture), meaning Blast (with multi-core encode and offload) can outperform it for continuous motion. This translates to potentially lower framerates or more frames dropped in full-motion scenarios under PCoIP (unless one uses GPU capture which VMware Horizon doesn’t support for PCoIP, only Blast). In short, PCoIP can be less efficient on server resources for heavy multimedia relative to Blast with NVENC.
• Limited Future in VMware Horizon: A practical drawback – VMware’s announcement means PCoIP will not receive new features in Horizon and will eventually not be available at all in Horizon after 2025 (Announcing end of support for PCoIP in VMware Horizon – VMware End-User Computing Blog). This makes PCoIP a “legacy” option in the VMware context. Organizations using Horizon will be pressured to move to Blast. If they absolutely need PCoIP, they may have to stay on older Horizon versions or switch to a different broker (which is a significant decision). This looming end-of-support diminishes confidence for some in continuing with PCoIP in VDI. HP Anyware will continue PCoIP development, but that’s a different product without the full VDI stack that VMware provides. Essentially, PCoIP’s innovation (apart from Ultra) might stagnate for VDI, focusing instead on niche remote workstation uses.
• Hardware Constraint for Zero Clients: Zero clients are a double-edged sword – while they are secure and simple, they are also inflexible. A PCoIP Zero Client cannot run productivity apps locally, cannot display content outside of a PCoIP session, etc. If an enterprise wants to, say, support both Horizon (Blast) and another system, zero clients might not work for the other system. Thin clients (with Windows or Linux) can have multiple clients (PCoIP, Blast, Citrix, etc.). So the zero client, if not updated or if needing replacement to support new protocols, becomes a bit of a dead-end investment. Many zero clients also cap out at certain video resolutions or codecs (TERA2 zero clients won’t ever support H.265, for instance). Organizations that have them love their low maintenance, but the cost to replace them for new needs is a consideration – this is a PCoIP-specific weakness in that Blast never required specialized client hardware. (However, note new Anyware clients from HP are introducing “zero clients” that can handle both PCoIP and Blast, which mitigates this).
• Ecosystem and Feature Catch-Up: Some newer remote experience enhancements have been focused on Blast (by VMware) and not implemented for PCoIP. For example, VMware integrated Microsoft Teams optimization with Blast using Media Optimization (offloading video calls to the client). In Horizon, if you use PCoIP, you might not get the same Teams optimization (or Zoom offload) – those are Blast-only. Additionally, things like HTML5 browser access, or client drive redirection enhancements, were prioritized for Blast. PCoIP in Horizon has been essentially in maintenance mode. So in a Horizon environment, using PCoIP could mean missing out on a few features or having less support for the latest innovations. Even outside VMware, one could argue the broader industry momentum (GPU encoding, etc.) is around H.264/HEVC. PCoIP is proprietary, so third-party tools (like monitoring, WAN accelerators) might not support it as well as they do standard protocols. This can be seen as a disadvantage in environments where ecosystem integration is important.
• Network Port Requirements: Minor, but worth noting – PCoIP needs open UDP 4172 (and TCP 4172) ports on firewalls or VPNs. In some secure environments, convincing the network team to allow that custom port can be a challenge, whereas Blast can piggyback on HTTPS (443). This is usually solvable (and Horizon’s Security Server/UAG can tunnel PCoIP over TCP 443 if needed, with some performance hit), but it’s a deployment consideration where Blast is simpler.

Blast Extreme – Strengths and Weaknesses

Benefits/Strengths:

• Broad Client & Device Support: Blast was designed for the widest range of client devices – from desktops to smartphones to browsers. Any device with an H.264 decoder can use Blast. This makes BYOD and multi-platform support much easier. Users can run their virtual desktop on an iPad during a commute, then on a Windows PC in the office, all using the same Blast protocol. Over 70 certified thin/zero client models support Blast Extreme natively as well (In detail — VMware’s Blast Extreme protocol%2C Linux)). The availability of an HTML5 option (Blast HTML Access) means even if a user can’t install a client, they can access via a web browser. PCoIP has no equivalent web access mode. This ubiquity is a huge advantage in modern IT environments that value flexibility.
• Excellent Video & Multimedia Performance: Because Blast leverages H.264/HEVC, which are video codecs highly optimized for motion, it delivers smooth video playback and 3D rendering. With GPU acceleration (NVENC or AMD VCE), Blast can stream even 4K video or high-fps content with low latency and without pegging the CPU (In detail — VMware’s Blast Extreme protocol) (In detail — VMware’s Blast Extreme protocol. It’s arguably better suited for multimedia-heavy use cases like training videos, YouTube playback, etc., especially on the client side – the client’s GPU decodes the stream with minimal power consumption (Understand the basics of VMware Blast Extreme vs. PCoIP | TechTarget) (VMware Horizon 7 Blast Extreme Primer—Everything an Admin Needs to Know – VMware End-User Computing Blog). Users notice that Blast handles full-screen video or GUI animations (like 3D transitions) with consistently high frame rates if bandwidth allows. Additionally, VMware integrated audio-video sync and other media streaming improvements into Blast to improve things like live video conferencing. For instance, Blast can leverage VMware Virtual Webcam and RTAV (Real-Time Audio-Video) optimizations for Zoom/Teams, something PCoIP struggled with in older implementations. In short, for a workforce that expects to watch videos or do Zoom calls in their virtual desktop, Blast provides a better experience out-of-the-box.
• Lower Client Resource Usage (Great for Mobile): One of VMware’s key design points was offloading decode to client hardware. On a modern laptop or mobile device, decoding Blast’s H.264 stream uses the dedicated video decoder, which means lower CPU usage and longer battery life on laptops/tablets (In detail — VMware’s Blast Extreme protocol) (Understand the basics of VMware Blast Extreme vs. PCoIP | TechTarget). PCoIP is decoded in software (except on zero clients) and can tax a client CPU, especially older ones. Blast clients thus tend to run cooler and more efficiently. A blog post comparing protocols noted “Blast Extreme had a much better user experience and also uses a lot LESS resources on the endpoint” (Remote protocols benchmarking, Citrix, VMware and RDP–Part One PCoIP vs Blast Extreme | Marius Sandbu – IT blog). This is beneficial for thin clients with modest CPUs – they can still render full HD Blast sessions smoothly because the heavy lifting is done by a small video decode ASIC. This advantage is amplified on mobile networks (where efficient compression means less radio usage and power). It makes Blast ideal for scenarios like users on laptops working over cellular hotspots – Blast’s efficiency helps preserve battery whereas PCoIP would drain it faster due to CPU work.
• Integration & Future Development: Blast is VMware’s own protocol, so it gets first-class integration with the whole Horizon suite and ongoing R&D investments. Features like VMware’s Blast Unity Touch (for mobile-friendly UI interactions), Blast clipboard, and optimized drives all tie into the Horizon agent. VMware also aligns Blast with upcoming tech – for example, as AV1 video codec emerges, VMware could incorporate it into Blast if beneficial. Customers benefit from a single vendor (VMware) handling end-to-end support for Blast issues. Also, VMware’s decision to consolidate on Blast means all enhancements (security updates, performance tuning, new features) will go there. Over time, this will likely widen any gap in capability in favor of Blast. Already, VMware has built a robust Blast Extreme Adaptive Transport (BEAT) that dynamically chooses UDP vs TCP based on network tests (Deep Dive into VMware Horizon Blast Extreme Adaptive Transport) – making it easier for admins (no need to choose manually). This continuous improvement trajectory is a strength, as Blast is not static; each Horizon release often brings Blast performance or feature tweaks, guided by VMware’s large install base feedback. Essentially, Blast is the strategic protocol for VMware/Broadcom, so it will evolve with virtualization trends (cloud, high-res monitors, etc.). Meanwhile, PCoIP (in Horizon) is frozen.
• Firewall/NAT Friendliness: Blast uses TCP 443/UDP 443 (or 8443) which are commonly open ports. In many cases, Blast traffic can pass through corporate firewalls, proxies (if configured for TLS tunneling), and NAT devices without special configuration – it looks like standard HTTPS. This simplifies remote access deployments. For example, using VMware Unified Access Gateway, all Blast traffic from the internet is on port 443 and can be demultiplexed by the gateway (VMware Horizon 7 Blast Extreme Primer—Everything an Admin Needs to Know – VMware End-User Computing Blog). This means end-users in hotels or cafes (which often only allow web traffic) can connect their virtual desktop without VPN just by using Blast over port 443. PCoIP by contrast often required VPN or special firewall rules for UDP 4172. The ease of access with Blast is a big plus for IT departments supporting work-from-anywhere users. It also means less troubleshooting “can’t connect” issues due to network middleboxes, since Blast can fallback to TCP 443 if UDP is blocked, ensuring connectivity (with some performance cost) in restrictive networks.
• Feature Parity and Extras: By now, Blast supports virtually all the features needed: multi-monitors, smart card auth, USB redirection, printer redirection, etc., matching PCoIP. Additionally, Blast introduced some unique features like “Blast Live Analytics” (with Horizon Performance Tracker, you can see Blast session details like encoding, packet loss, etc.). Admins can fine-tune Blast with a rich set of policy settings (codec selection, frame rate caps, image quality levels). Blast also supports HDR10 color in recent versions (allowing higher color depth for HDR content on supported clients), which PCoIP doesn’t mention support for. These little extras may not be widely used yet, but they show Blast is keeping up with modern demands (e.g., HDR, high DPI scaling awareness, etc.). Blast’s compatibility with vGPU, including VMware’s support for VMware vSphere vGPU live migration (which PCoIP doesn’t directly control), means power users on Blast can even have HA failover now without losing session (vSphere can checkpoint a GPU VM and restore – this was tested with Blast). So, in enterprise environments requiring high availability, Blast plays nicely.
• Reduced Server Load with Offloads: Using technologies like NVENC and Intel Quick Sync, Blast can drastically reduce VDI server CPU usage. A GPU like NVIDIA Tesla M10 can encode many Blast sessions in parallel. VMware noted up to 80% bandwidth reduction and 50 ms latency reduction using Blast with NVENC vs PCoIP (In detail — VMware’s Blast Extreme protocol (In detail — VMware’s Blast Extreme protocol). This means higher user density per server when GPUs are present – a cost benefit. Even with CPU encode, Blast can leverage multiple cores (the Blast encoder in Horizon is multi-threaded, unlike Horizon’s PCoIP which was single-thread limited to SSE2 optimizations (PCoIP Ultra Technical Series Part 1: Top-Level Architecture)). So Blast can scale with modern multi-core servers better, potentially encoding faster or multiple displays concurrently more efficiently. This scalability translates to supporting more concurrent users or higher resolutions per host.

Drawbacks/Weaknesses:

• Potentially Higher Bandwidth for Sharp Text or Graphical Fidelity: Because Blast primarily uses lossy video compression, it doesn’t always preserve fine details unless more bandwidth is spent or lossless mode is enabled. By default, Blast will sacrifice some fidelity to save bandwidth – e.g. small fonts or thin lines might look slightly blurry due to H.264 compression at low bitrates. Achieving true lossless in Blast is possible but “compromises frame rates and raises CPU or bandwidth usage” () (). PCoIP’s design of sending refined lossless layers is more bandwidth-thrifty for those details. If an organization cares deeply about, say, exact color precision or perfect image quality but also minimizing bandwidth, Blast might disappoint because H.264 (even 4:4:4 mode) cannot recover missing information without raising bitrate significantly () (). In practice, Blast’s default profile is tuned to be very good but not mathematically perfect. So one could label a Blast weakness as visual fidelity for certain content types – especially comparison side by side with PCoIP: e.g., a small 1px checkerboard pattern might smear under H.264 unless lossless is on. For most users this is negligible, but for power users (artists, engineers), it’s a consideration. Also, as noted earlier, if Blast is configured for max fidelity (e.g. lossless every frame), it will consume a lot of bandwidth, much more than PCoIP’s selective lossless approach (Blast Extreme Requires GPUs to Deliver a Competitive User Experience) (Blast Extreme Requires GPUs to Deliver a Competitive User Experience). Thus, while Blast can be tuned for quality or for low bandwidth, it can’t have both as elegantly as PCoIP’s progressive build. In bandwidth-constrained situations, Blast might end up showing more compression artifacts than PCoIP, or use more data to avoid them.
• Reliance on Hardware for Best Results: Blast’s strong performance is partly because it taps into hardware encoders/decoders. If that hardware isn’t present (e.g. no GPU on server, or an older thin client without H.264 offload), then the burden falls to software and can become a bottleneck. A Horizon deployment with no GPUs but using Blast for, say, 3D CAD will see very high CPU usage per session, limiting scalability (20% more CPU than PCoIP per session was observed in one test) (Blast Extreme Requires GPUs to Deliver a Competitive User Experience). Similarly, if a user is on an old PC without a good GPU, decoding a 4K Blast stream might use a lot of their CPU, whereas PCoIP’s adaptive approach might have sent a lighter stream. So Blast is optimized for modern hardware – which generally is available, but not always in every scenario. If hardware encoding fails or is misconfigured, Blast could perform worse. With PCoIP, performance is more consistent regardless of hardware (since it never expected special hardware except the zero client itself). Also, Blast’s multi-monitor with high resolution works best with GPU (because CPU encoding 4x 4K displays is nearly impossible in realtime). PCoIP hardware (multiple chips) could handle that in some implementations. So one could say Blast scales down less gracefully in environments without acceleration or with legacy gear.
• Network Constraints and TCP Fallback Issues: While Blast has adaptive transport, if network admins, for example, don’t open UDP and Blast is forced into TCP mode on a latent network, performance suffers. Some early criticisms of Blast (circa 2016) were that it used more bandwidth and had worse lag on lossy links when stuck on TCP (Remote protocols benchmarking, Citrix, VMware and RDP–Part One PCoIP vs Blast Extreme | Marius Sandbu – IT blog). VMware has mitigated this, but the fact remains that Blast’s performance is highly dependent on network configuration – it ideally wants UDP allowed. PCoIP being UDP from the start meant everyone deploying it had to accommodate that (often via VPN). With Blast, there might be deployments where admins leave it on TCP (for simplicity or ignorance of best practice), and users might experience sluggishness in high-loss or high-latency conditions. So a “weakness” is that Blast can still be sensitive to network configuration – it needs tuning (like enabling UDP, adjusting MTU, etc.) to truly match PCoIP’s robustness. If left untuned, PCoIP might outperform it on challenging networks simply because it was always UDP and adaptive.
• Vendor Lock-in and Licensing: Blast is proprietary to VMware (Broadcom). This means if you want to use Blast, you have to use VMware Horizon (or derivatives like Horizon Cloud). There’s no third-party Blast broker or Blast client outside VMware’s ecosystem. This lock-in can be viewed as a drawback for those who want multi-cloud or hybrid solutions – you can’t, for instance, use Blast to connect to an AWS WorkSpaces instance or a physical workstation unless it’s running Horizon agent. PCoIP, on the other hand, is available in multiple platforms (VMware, AWS, HP Anyware, etc.). Moreover, if a customer is unhappy with VMware licensing costs or support, they cannot “take their Blast protocol” elsewhere easily; they’d have to switch protocols. Blast is included with Horizon (so not an extra cost item itself), but Horizon’s cost may be high and tied to it. Some organizations prefer not to rely on a single vendor for a core technology like remoting – in that sense, PCoIP (with multiple vendors supporting it) offered more flexibility. With Broadcom’s acquisition, some customers worry about potential price increases or changes in support model for Horizon/Blast. In contrast, PCoIP being available via HP might offer alternative licensing models (HP Anyware subscription could be more cost-effective for certain use counts). So a Blast user is essentially tied to VMware’s roadmap and pricing – if VMware/Broadcom decide on changes, the customer has limited recourse except migrating off Horizon entirely.
• Complexity in Tuning for Edge Cases: Because Blast has many codec options and configurable parameters, getting the absolute best performance for a given scenario might require expert tuning. For example, VMware published a 40-page “Blast Extreme Optimization Guide” with dozens of tips (codec settings for different network types, enabling “Latency Floor” or not, etc.) (40 Tips for Optimizing VMware Horizon VDI’s Blast Extreme Protocol) (Deep Dive into VMware Blast Extreme Network Intelligent Transport). This shows that while Blast can be adapted to many situations, it’s not always plug-and-play for atypical cases. Administrators may need to experiment with settings (e.g., increase the Blast Max Frame Rate or adjust the H.264 quality factor) to achieve desired results. PCoIP, conversely, had fewer knobs for admins – it auto-adapts quite well, with only broad profiles to set. So some might view Blast as a bit more complex to fine-tune. If misconfigured (say an admin fixes the quality too high on a low bandwidth link), Blast could perform poorly. Essentially, Blast’s flexibility comes at the cost of more configuration options, which can be a downside in simpler deployments.
• Certain Features Lag PCoIP (in non-VMware contexts): This is a minor note, but PCoIP in HP Anyware context supports some features like host-to-client EDM (accelerated media streaming) that Blast doesn’t have a concept of. Also, if looking outside VDI, PCoIP can do things like directly broker without a server via IP connection (point-to-point), which Blast cannot (Horizon is required). So if you wanted a quick ad-hoc remote session without full infrastructure, Blast isn’t an option whereas Teradici has lightweight brokerless options for PCoIP. These aren’t typical enterprise needs, but they show a limitation in Blast’s versatility outside its primary domain.

In summary, Blast Extreme’s strengths lie in its modern codec efficiency, wide accessibility, and alignment with current and future tech (making it ideal for most enterprise VDI deployments), whereas its weaknesses are mostly about needing current hardware, using lossy compression by default, and being tied to VMware’s ecosystem. PCoIP’s strengths are in its refined and balanced approach to quality vs. bandwidth and its platform neutrality, with weaknesses in upcoming support and multimedia efficiency without offload. The right choice can depend on the specific priorities of an organization (quality vs. cost vs. future-proofing vs. existing investments).

6. Visual Aids

To illustrate some of the differences discussed, below are a few charts and diagrams highlighting protocol architecture, performance metrics, and industry trends.

(VMware Horizon 7 Blast Extreme Primer—Everything an Admin Needs to Know – VMware End-User Computing Blog) Architecture and Connectivity: Example network architecture for VMware Blast Extreme in a Horizon deployment. Blast sessions can be brokered through a Security Server or Unified Access Gateway in the DMZ, using TCP/UDP 8443 (or TCP 443) for streaming. The diagram shows how Blast traffic (orange lines) and other services flow between client, DMZ, and data center components. A single secure port can carry Blast’s display, input, and peripheral traffic (VMware Horizon 7 Blast Extreme Primer—Everything an Admin Needs to Know – VMware End-User Computing Blog) (VMware Horizon 7 Blast Extreme Primer—Everything an Admin Needs to Know – VMware End-User Computing Blog), simplifying firewall configuration compared to PCoIP’s dedicated port approach. (Source: VMware)

(Blast Extreme Requires GPUs to Deliver a Competitive User Experience) Relative Bandwidth Efficiency: A Teradici benchmark comparing PCoIP vs. Blast Extreme (H.264) for an office productivity workload. Lower bars mean less network bandwidth used. Blue bars are PCoIP (software encode on host, with and without a Hardware Accelerator), green bars are Blast H.264 (software vs. GPU NVENC, and TCP vs. UDP transport). The chart indicates PCoIP generally used less bandwidth than Blast for the same task when both were software-encoded (blue vs. green on left) (Blast Extreme Requires GPUs to Deliver a Competitive User Experience). Blast with GPU (far right green bars) increased bandwidth substantially (to maintain quality), whereas PCoIP’s bandwidth stayed modest even with hardware assist (blue bar second from left) (Blast Extreme Requires GPUs to Deliver a Competitive User Experience) (Blast Extreme Requires GPUs to Deliver a Competitive User Experience). This visual reinforces how PCoIP is tuned for bandwidth conservation, while Blast tends to use more bandwidth when more performance headroom (GPU/UDP) is available.

(Adoption Trend Note: According to industry reports and vendor statements, by 2020–2021 a majority of VMware Horizon deployments had shifted to Blast as the primary protocol, while PCoIP remained strong in AWS WorkSpaces and specialized on-prem use. The announced PCoIP support sunset in Horizon by 2025 (Announcing end of support for PCoIP in VMware Horizon – VMware End-User Computing Blog) underscores this trend. HP’s acquisition of Teradici in 2021 and their continued development of PCoIP (Ultra, Anyware) indicate that PCoIP will persist and evolve outside VMware’s sphere, targeting new markets like remote work solutions independent of VDI suites. Over the next few years, we can expect Blast Extreme to cement its dominance in the VMware user base, whereas PCoIP could see renewed adoption via HP in media/entertainment and OEM partnerships. In essence, each protocol is finding its distinct niche in the market.)

7. Conclusion

Both PCoIP and Blast Extreme are high-performance remote display protocols that have significantly advanced the VDI and remote work landscape. PCoIP brought unparalleled image quality and a “local-like” experience to virtual desktops, setting a gold standard for user experience through the late 2000s and 2010s. Its innovative progressive codec and secure, stateless client model made it a favorite for security-sensitive and graphics-intensive applications. Blast Extreme, developed a few years later, built on industry-standard codecs to achieve similar excellence while broadening accessibility and catering to modern client devices and GPUs. Over time, Blast has matured to offer comparable, and sometimes superior, performance and features, leading VMware to adopt it as the default and phase out PCoIP.

For IT professionals, the choice between PCoIP and Blast may not be as simple as “which is better,” but rather “which is better for our use case and environment.” If an organization is deeply invested in VMware’s ecosystem and wants a future-proof solution with broad device support, Blast Extreme is the clear direction. It will continue to get enhancements and is already delivering great results in most deployments, especially where video, unified communications, and mobile access are important. On the other hand, if an organization values the ultimate in image fidelity, has existing Teradici endpoints, or needs a platform-agnostic protocol (for example, to cover multi-cloud or niche workstation remoting), PCoIP (via HP Anyware) remains a compelling choice – it offers a proven track record and ongoing improvements like PCoIP Ultra for modern workloads.

In many environments, a hybrid approach may even be used: for instance, running Blast for general office staff VDI, but using PCoIP for a subset of power users or contractors via HP Anyware. As of this writing, VMware Horizon even allows per-pool protocol selection, so some pools could use PCoIP (perhaps legacy reasons or specific needs) while others use Blast – giving flexibility during transitions.

From a cost perspective, organizations should weigh the licensing implications. VMware Horizon (with Blast) consolidates costs under one vendor. Switching entirely to HP Anyware (PCoIP) would introduce a separate licensing cost but could eliminate some VMware costs if one foregoes Horizon – however, HP Anyware is not a full VDI stack (no integrated broker for multi-session farms, etc., it often complements rather than replaces Horizon). Many will find staying with VMware Blast the simpler and more supportable route, as VMware’s support will focus on Blast issues.

In terms of user experience, both protocols can provide an excellent experience when properly configured. User feedback often indicates they can’t tell much difference during normal tasks if bandwidth is ample and clients are capable. Edge cases will highlight differences (Blast might pixelate a little during a sudden fast motion on a weak network; PCoIP might reduce frame rate on the same; PCoIP text might be sharper; Blast video smoother, etc.). With prudent tuning, these can be mitigated.

Scalability and reliability in large deployments have been proven for both. PCoIP has huge production deployments (e.g. tens of thousands of AWS WorkSpaces users daily, large government installations) and so does Blast (major banks, universities, and enterprises with global Horizon deployments). Both protocols can be scaled horizontally by adding more connection servers or gateways as needed. In terms of reliability, both use mature technologies – one might argue PCoIP’s simpler UDP design gives one less thing to go wrong (no “fallback to TCP” scenarios), whereas Blast’s reliance on network conditions means it needs a bit more care (ensuring UDP paths, etc.).

In conclusion, PCoIP and Blast Extreme each have strengths that make them suitable for different scenarios, and an IT professional’s task is to map those to their organization’s requirements. PCoIP’s legacy and continued specialty use underscore its strength in quality and security, while Blast’s momentum and all-round capabilities make it the go-to for most new VDI deployments. As the industry stands in 2025, Blast Extreme is becoming the de facto standard within VMware environments, and PCoIP is evolving in the hands of HP for specialized use. Keeping an eye on future developments (such as new codec integrations like AV1, or further optimizations in HP Anyware) will be important. Ultimately, either protocol – when implemented with best practices – can deliver a secure, efficient, and high-performance remote desktop experience for users, fulfilling the promise of being able to work anywhere on any device without compromising on productivity.

References:

1. Teradici corporation background and PCoIP history (Teradici – Wikipedia) (Teradici – Wikipedia)
2. VMware’s adoption of PCoIP in View 4.0 (2009) (Teradici – Wikipedia) (A complete history of VMware and VDI)
3. VMware Blast introduction and evolution (Announcing end of support for PCoIP in VMware Horizon – VMware End-User Computing Blog) (Understand the basics of VMware Blast Extreme vs. PCoIP | TechTarget)
4. PCoIP technical architecture and benefits (Teradici/HP Anyware documentation) (About PCoIP Technology – HP Anyware Architecture Guide) (About PCoIP Technology – HP Anyware Architecture Guide)
5. PCoIP host rendering, progressive build-to-lossless and network adaptation (Microsoft Word – TER0806005 Issue 1 – WP-Network Infrastructure Requirements_FINAL_July4_bhz.doc) (Microsoft Word – TER0806005 Issue 1 – WP-Network Infrastructure Requirements_FINAL_July4_bhz.doc)
6. Blast Extreme technical details (VMware blog and Bas van Kaam analysis) (In detail — VMware’s Blast Extreme protocol) (In detail — VMware’s Blast Extreme protocol)
7. Performance comparisons between Blast and PCoIP (Teradici and VMware benchmarks) (Blast Extreme Requires GPUs to Deliver a Competitive User Experience) (Blast Extreme Requires GPUs to Deliver a Competitive User Experience)
8. Blast vs PCoIP bandwidth and latency observations (Remote protocols benchmarking, Citrix, VMware and RDP–Part One PCoIP vs Blast Extreme | Marius Sandbu – IT blog) (VMware Horizon Blast Extreme Acceleration with NVIDIA GRID | VMware End-User Computing Blog)
9. Security features of PCoIP (AWS and EVGA documentation) (Data protection in Amazon WorkSpaces – Amazon WorkSpaces) (Microsoft Word – EVGA PCoIP_User_Guide.doc)
10. VMware announcement on ending PCoIP support in Horizon (Announcing end of support for PCoIP in VMware Horizon – VMware End-User Computing Blog) (Announcing end of support for PCoIP in VMware Horizon – VMware End-User Computing Blog)
11. ClearCube whitepaper on PCoIP vs H.264 protocols () ()
12. Bas van Kaam’s “Inside Blast Extreme” detailing codec and transport aspects (In detail — VMware’s Blast Extreme protocol%2C Linux)) (In detail — VMware’s Blast Extreme protocol)
13. Teradici blog on Blast vs PCoIP (performance and cost analysis) (Blast Extreme Requires GPUs to Deliver a Competitive User Experience) (Blast Extreme Requires GPUs to Deliver a Competitive User Experience)
14. TechTarget article on Blast Extreme vs PCoIP use-cases (Understand the basics of VMware Blast Extreme vs. PCoIP | TechTarget) (Understand the basics of VMware Blast Extreme vs. PCoIP | TechTarget)
15. AWS WorkSpaces documentation on PCoIP encryption and ports (Data protection in Amazon WorkSpaces – Amazon WorkSpaces) (Data protection in Amazon WorkSpaces – Amazon WorkSpaces)
16. VMware Blast Extreme primer (technical whitepaper) (VMware Horizon 7 Blast Extreme Primer—Everything an Admin Needs to Know – VMware End-User Computing Blog) (VMware Horizon 7 Blast Extreme Primer—Everything an Admin Needs to Know – VMware End-User Computing Blog)
17. Teradici PCoIP Ultra architecture blog (PCoIP Ultra Technical Series Part 1: Top-Level Architecture) (PCoIP Ultra Technical Series Part 1: Top-Level Architecture)
18. Reddit user experiences comparing Blast and PCoIP (Remote protocols benchmarking, Citrix, VMware and RDP–Part One PCoIP vs Blast Extreme | Marius Sandbu – IT blog)
19. VMware Horizon Blast with NVIDIA GRID performance results (VMware Horizon Blast Extreme Acceleration with NVIDIA GRID | VMware End-User Computing Blog) (VMware Horizon Blast Extreme Acceleration with NVIDIA GRID | VMware End-User Computing Blog)
20. Teradici claims on Blast requiring GPUs for parity (Blast Extreme Requires GPUs to Deliver a Competitive User Experience) and PCoIP bandwidth advantages (Blast Extreme Requires GPUs to Deliver a Competitive User Experience) (Blast Extreme Requires GPUs to Deliver a Competitive User Experience).