• The Four Hundred
  • Subscribe
  • Media Kit
  • Contributors
  • About Us
  • Contact
Menu
  • The Four Hundred
  • Subscribe
  • Media Kit
  • Contributors
  • About Us
  • Contact
  • Expand’s New Data Acceleration Technology Proves Fast in Tests

    November 16, 2004 Alex Woodie

    Researchers at Expand Networks have come up with a new WAN compression algorithm called Data Transfer Acceleration (DTA) that can achieve compression rates of up to 20-1. That figure is the result of a test that Expand commissioned Tolly Group to perform against the company’s two main competitors for WAN compression appliances, Packeteer and Peribit, the results of which were announced yesterday.

    Expand paid Tolly, an independent organization that specializes in testing network equipment, to benchmark a pair of Expand Accelerator 4820 devices, running at Version 5.03, against two Packeteer PacketShaper 2500 Version 6.2.1 devices and two Perebit Sequence Reducer-20 Version 5.0.1.4 devices (these devices are always deployed in pairs). All three vendors use similar technologies and techniques in their WAN compression appliances to get more bandwidth out of a given Internet link by eliminating redundancies in IP datastreams, although there are some differences among them in feature sets.

    Expand’s Accelerators delivered 13 times the effective WAN link capacity when compared with file transfers without compression, according to the test by Tolly. Expand’s appliances provided nearly eight times more WAN link capacity than the other devices, and delivered file transfer times that were five times faster than the Peribit and Packeteer products.

    Expand’s chief executive, Ariel Shulman, says the test results validate the work the company put into researching and developing its new DTA algorithm, which was provided as a minor update to its AcceleratorOS Version 5.0, released earlier this year. “This is the DTA algorithm, which outperforms our previous algorithm, and clearly our competitors’ algorithms, as you can see from the graph,” Shulman says. “We were roughly as good as Packeteer before.”

    Not surprisingly, neither Peribit nor Packeteer participated in Tolly’s test, although they were asked to. Jeff Barker, a director of product management at Packeteer, says the PacketShaper appliance definitely would have performed better in the tests had the Tolly Group used the latest Version 7 release of Packeteer’s software, which it recently released. Perebit had not responded to requests for comment by this newsletter at press time.

    THE TOLLY TESTS

    Tolly ran a series of tests to measure both the rate of compression and the total file transfer times achieved by the WAN compression devices. The data sent across the links was the Canterbury Files, a collection of large data files that is the de-facto standard for testing data compression algorithms. The Canterbury Files, developed by the CS department at the University of Canterbury in New Zealand, include the Bible, the first million digits of Pi, the complete genome of the E. Coli bacteria, and other large and varied data sets.

    None of the devices fared very well in the first round of the compression test, which is expected for these types of devices, as they must first be exposed to the IP traffic so they can “learn” where the redundancies are. In this first round, the Expand device enabled the WAN link to deliver 156 percent of its standard bandwidth, the Perebit device 142 percent, and the Packeteer device 128 percent–which is to say, the devices increased the network bandwidth by anywhere from about one quarter to about one half, which isn’t much for a device that starts at $4,500, in the case of the Expand Accelerator 4820.

    The second round of compression tests was a different story. On average, the Expand device was able to fill up the WAN link (set up to simulate a 512 Kpbs frame relay line) to more than 1,200 percent of its stated capacity, with several of the files being compressed at or near a 20:1 ratio, which translates into using 2,000 percent of the WAN’s stated bandwidth capacity. The Packeteer device achieved an average of 545 percent utilization of the WAN, while the Perebit device had an average of 196 percent. The compressibility of the files varied greatly, no matter which device was doing the compressing, and the Packeteer exhibited 15:1 compression on two files, and 10:1 one on file. The Perebit device compressed two files at or near the 5:1 ratio, according to the results.

    Because of network latency and other issues, the capability to compress files doesn’t always translate directly into faster file transfer speeds, which is why Tolly conducted a second set of tests to test file transfer times across three different types of WANs, including simulation of a high-latency 128 Kbps satellite link, a medium-latency 512 Kpbs “long-haul” frame relay link, and a low-latency 2 Mbps T-1 line.

    On the slowest satellite link, it took nearly 200 seconds to transfer a human DNA file without any compression. With the Expand devices plugged in, the same file was transferred three times faster, in about 65 seconds, while the Packeteer and Perebit devices took 163 seconds and 127 seconds, respectively. The Expand device outperformed the other two devices in every other speed test transferring other Canterbury Files across the three WAN scenarios, according to the test results.


    The results show the winner was Expand, says Kevin Tolly, president and chief executive of Tolly Group, which is based in Boca Raton, Florida. “We stress-tested all three devices under multiple WAN scenarios to simulate real-world production networks,” Tolly says. “The Expand Accelerator 4820 outperformed the competition in every test scenario, demonstrating that the device enables users to maximize their WAN connections efficiently.”

    Although the test results showed the Expand Accelerator appliances can enable a given WAN link to deliver up to 2,000 percent of its stated bandwidth (a 20:1 compression ratio), Shulman says his company will maintain its previous policy of claiming an average improvement of 100 to 400 percent, “with peaks of 1,000 percent.” “We recently added ‘peaks of 1,000 percent or more,’ ” Shulman says. “We still like to under-promise and over-deliver.”

    The new ExpandOS Version 5 and DTA algorithm will be available to OS/400 shops that bought an OEMed version of the Expand Accelerators through Lakeview Technology, which became an Expand partner earlier this year and is reselling a modified version of the Accelerator it calls MIMIX Network1.

    Share this:

    • Reddit
    • Facebook
    • LinkedIn
    • Twitter
    • Email

    Tags:

    Sponsored by
    Maxava

    Migrate IBM i with Confidence

    Tired of costly and risky migrations? Maxava Migrate Live minimizes disruption with seamless transitions. Upgrading to Power10 or cloud hosted system, Maxava has you covered!

    Learn More

    Share this:

    • Reddit
    • Facebook
    • LinkedIn
    • Twitter
    • Email

    Mad Dog 21/21: Vow Tipping Discover the DBVisualizer

    Leave a Reply Cancel reply

Volume 4, Number 46 -- November 16, 2004
THIS ISSUE
SPONSORED BY:

LANSA
AURA Equipments
iTera
Bytware
Affirmative Computer

Table of Contents

  • PGM Provides Springboard for iSeries Web Development
  • Startup Delivers Database Tools for OS/400 HA Environments
  • Expand’s New Data Acceleration Technology Proves Fast in Tests
  • Linoma Offers PGP Studio with Data Transfer Tool

Content archive

  • The Four Hundred
  • Four Hundred Stuff
  • Four Hundred Guru

Recent Posts

  • Public Preview For Watson Code Assistant for i Available Soon
  • COMMON Youth Movement Continues at POWERUp 2025
  • IBM Preserves Memory Investments Across Power10 And Power11
  • Eradani Uses AI For New EDI And API Service
  • Picking Apart IBM’s $150 Billion In US Manufacturing And R&D
  • FAX/400 And CICS For i Are Dead. What Will IBM Kill Next?
  • Fresche Overhauls X-Analysis With Web UI, AI Smarts
  • Is It Time To Add The Rust Programming Language To IBM i?
  • Is IBM Going To Raise Prices On Power10 Expert Care?
  • IBM i PTF Guide, Volume 27, Number 20

Subscribe

To get news from IT Jungle sent to your inbox every week, subscribe to our newsletter.

Pages

  • About Us
  • Contact
  • Contributors
  • Four Hundred Monitor
  • IBM i PTF Guide
  • Media Kit
  • Subscribe

Search

Copyright © 2025 IT Jungle