Peribit Networks introduced an Application Flow Acceleration (AppFlow) technology that integrates compression and caching, acceleration, application control, and visibility technologies to improve the performance of applications over the WAN. The technology transparently speeds performance of Microsoft Exchange, file services, and web-based applications over point-to-point links.
Peribit said its new AppFlow technology provides the acceleration needed for applications that cannot benefit from TCP acceleration until their own protocol limitations are removed. Microsoft Exchange and Common Internet File System (CIFS) file services applications were initially designed for local area network (LAN) environments, so their performance drops dramatically when used across a WAN link that has even modest latency. Similarly, web pages for in-house and packaged applications build very slowly across the WAN because they depend on sequential delivery of tens to hundreds of objects.
Peribit’s AppFlow technology accelerates the performance of applications based on three specific protocols: the Messaging Application Programming Interface (MAPI) used by Exchange; Microsoft CIFS; and HTTP. The AppFlow technology improves response times for remote Outlook/Exchange users by requesting portions of the message in advance of the client’s request, pipelining what had been a serial and inefficient bulk transfer. As a result, the entire e-mail and any attachments are local to the client by the time the client requests the message. Peribit Networks’ ability to accelerate Exchange benefits even fairly low-latency links of 30 ms or more. Similar performance gains are seen for the CIFS protocol. To speed web-based application performance across the WAN, Peribit devices learn the objects associated with a URL and pipeline them across the link so that they are available locally when the page is requested a second time.
Peribit also announced two new additions to its Sequence Mirror product family, which uses sequence caching technology to record patterns across packets, sessions, and flows and store them on hard disk so that, if repeated later, they can be identified and removed from data streams even when they are separated by hundreds of gigabytes of data sent days or weeks apart, and even when the transmitted files have been modified. http://www.peribit.com
