Evaluating Citrix Branch Repeater for WAN optimization

Citrix's product for WAN optimization is the Citrix Branch Repeater. This family of products integrates network optimization with application, desktop and server virtualization.

Citrix's heritage is that of a software company focused on providing thin-client solutions. It was so successful

that many people refer to a class of application traffic as "Citrix" when they mean traffic that originates from Citrix's XenApp (formerly Citrix Presentation Server ) using the Independent Computing Architecture (ICA) protocol. Citrix currently focuses on application delivery and is focused on the emerging desktop-virtualization market with its XenDesktop product.  The company states that its goal is to help IT organizations transform static data centers into dynamic delivery centers where applications and desktops are centrally managed and delivered on demand.

Citrix's WAN optimization products consist of the Citrix Branch Repeater family (formerly known as WANScaler), which includes the Branch Repeater and Branch Repeater with Windows Server appliances for branch offices, Repeater appliances for data centers and larger sites, and a software client (called the Repeater Plug-in for Citrix Receiver) for mobile users. In May 2009, Citrix publicly announced plans to deliver a fully software-based Branch Repeater virtual appliance in the first half of 2010.

Citrix has strategic partnerships with HP, Dell, CSC, Amazon and Microsoft, among others. Citrix claims that because of its relationship with Microsoft that Citrix branch office solutions natively integrate essential Windows branch services and management with accelerated delivery of Microsoft applications. Citrix also says that because of this relationship, it has access to Microsoft protocol specifications and this allows Citrix to develop robust and highly performing solutions for Microsoft applications such as File Sharing (CIFS) and Exchange (MAPI).

Important facts about the Citrix Branch Repeater

Citrix's Branch Repeater offers priority queuing based on user-configured application bandwidth limits. IT organizations can create and manage up to 64 service classes based on port numbers, specific IP address or IP subnet. IT organizations assign quality of service (QoS) priority to these service classes based on their business needs. Bandwidth utilized by lower priority classes of traffic is controlled to allow as much bandwidth as is needed by higher priority classes.

This priority-queuing QoS engine is automatically enabled when the appliance detects application traffic congestion and protects the bandwidth reserved for business-critical applications. In the absence of application congestion, Branch Repeater reverts to a weighted-fair queuing scheme, where all application traffic receives equal access to the bandwidth. Branch Repeater's QoS implementation does not tag traffic and is fully compatible with network-level QoS tagging including MPLS (Multi-protocol Label Switching).
The Branch Repeater dynamically uses five different engines to compress data at the byte, pattern, and data-chunk level using hardware cache memory algorithms, pattern-level software compression and data-chunk matching of up to several hundred megabytes per compressed token. First-pass compression algorithms include ZLIB and LZS. Second-pass traffic uses the tokenizing engines in zero-latency cache, ultra-fast DDR2 RAM, and fast-access hard drives. Branch Repeater compression is application-aware for ICA, CIFS, MAPI, HTTP, FTP and NFS in order to separate any embedded headers or messages from the data payload and maximize compression ratios.

The TCP optimization techniques provided by Branch Repeater include:

  • Optimizing bandwidth delay product (BDP) and TCP window sizes dynamically for all networks including long distance links, variable quality links (such as Internet VPNs) or multi-path networks.
  • Improving standard TCP congestion control mechanisms and their response to dropped packets.
  • Reducing the number of round trips it takes to complete a data transaction by compressing TCP payloads.

Citrix's Branch Repeater provides application-specific acceleration for ICA, CIFS, MAPI, NFS, HTTP and FTP. Citrix has announced that Secure Sockets Layer (SSL) acceleration is planned for a future release of the product.

Relative to its optimization of ICA, Citrix highlights the fact that ICA is designed to be efficient over any link and thus it is difficult to optimize ICA without directly orchestrating with the XenApp server. Branch Repeater uses a specific protocol optimization engine that participates in the ICA session and coordinates with the XenApp server for each user. To accelerate a user's session, Branch Repeater takes over the network transport, offloads compression, de-duplicates frequently accessed bandwidth-intensive data and graphics, and even decrypts and encrypts the ICA session. The ICA protocol acceleration leverages existing compression engines, and the algorithms are specifically tuned for interactive (i.e. mouse/keyboard and screen refresh) and bulk (i.e. printing over ICA) traffic flows.

For more, read all the sections in our guide:

About the author:
Dr. Jim Metzler, principal at Ashton Metzler and Associates, is a widely recognized authority on network technology and its business applications. In more than 28 years of experience, Jim has helped numerous vendors refine product and service strategies and has helped enterprises evolve network infrastructures. He has directed and conducted market research at a major industry analyst firm and run a consulting firm. Jim holds a Ph.D. in numerical analysis from Boston University. He is co-author of the book Layer 3 Switching: A Guide for IT Professionals (Prentice Hall).

This was first published in February 2010

Dig deeper on Bandwidth and capacity planning

Pro+

Features

Enjoy the benefits of Pro+ membership, learn more and join.

0 comments

Oldest 

Forgot Password?

No problem! Submit your e-mail address below. We'll send you an email containing your password.

Your password has been sent to:

SearchNetworking

SearchUnifiedCommunications

SearchTelecom

SearchSDN

Close