When a major company is about to announce a new product, service or an enhancement to a product, I often get many pre-emptive email messages from PR people who are hoping to co-opt the larger company's launch messages by providing commentary before the fact. Quite often these are thinly disguised marketing pitches.
From time to time, however, I receive a message that not only serves that purpose, but goes on to provide a useful bit of insight. I thought I'd share part of such a message here.
Next week, Microsoft will release improvements to its Remote Desktop Protocol (RDP), a viewing protocol for desktop virtualization, that will better support multimedia, streaming audio and rich media applications.
It’s a welcome improvement, but it doesn’t do much to help the IT managers who have to slog through a virtual soup of viewing protocols, none of which is a silver bullet. Every major vendor has its own viewing protocol, trumpeted by each as the ideal solution. But each has advantages and disadvantages. In truth, there is no clear, decisive winner among the current vendor offerings.
No single viewing protocol for virtual desktop environments delivers a comprehensive answer. Instead, they offer - at best - two of the following three things:
- Low bandwidth requirements
- Low computational requirements and high scalability
- High quality end-user experience with responsive multimedia
Put another way, if you want a high-quality end user experience you are going to have to pay in terms of bandwidth consumption, processing power, or both.
Eric Hanselman, CTO for Leostream, believes the best course for large organizations is to keep options open and not get locked into a single viewing protocol.
Low-end users, for example, might be served well by RDP. High-end users with bandwidth to spare may be best-served by PC over IP. Then again, high-end users with bandwidth constraints may need EOP extensions to RDP. According to Hanselman, a flexible approach also enables organization to take advantage of new, emerging protocols, such as Red Hat’s much anticipated Spice, set to be unveiled in the coming months.
He can explain the advantages and disadvantages of a variety of viewing protocols, such as:
- Windows’ RDP: Because it’s bundled into Windows machines, it’s ubiquitous. But many users will find that, though it’s easy on bandwidth and processor cycles, the end-user experience leaves much to be desired.
- Teradici / VMware’s PC over IP: Designed originally for high performance graphics applications, it originally employed dedicated hardware for compression and encryption of desktop display streams. Now in its software-only version, it does a beautiful job with high-quality graphics. But, because it no longer has a dedicated chip to compress video efficiently, Software PC over IP requires a great deal of bandwidth.
- Citrix’ HDX: Because this runs over Citrix’ legacy ICA protocol, it could hold an answer for a vast following of Citrix shops. But there’s a catch: for good performance, HDX requires a specialized nVidia graphics card. This requirement adversely impacts scalability.
- NoMachine’s NX: A high performance viewing protocol geared for Unix and Linux desktops, it has strong graphics performance and is particularly well suited to challenging network environments. However, it does not provide the kinds of multi-media redirection that RDP v7 can provide, and it requires a client. It’s a compromise that does not tax bandwidth or processing excessively, but it also has some shortcomings in the user experience.
Leostream produces the industry’s most widely deployed vendor-independent Connection Broker, used in more than 220 deployments. In the course of developing Connection Broker deployments, Leostream routinely works with customers to determine which protocol(s) will best match the needs of their end users. This “real world” expertise, as well as familiarity with heterogeneous environments and large scale projects, puts Hanselman in a unique position to address the current protocol wars.
Well done Leostream! Your message was both interesting and informative. I do agree that if an organization has a diverse set of access devices and requirements, a connection broker can be a good fried. If your organization falls into that category, it might be worth looking into what Leostream and other competitors, such as HP, are doing.