HP
Takes The Lead In The Fast Ethernet Analyzer Race
The NetXRayPro/100S provides Host Table and Matrix graphs at the IP level. It also provides a large selection of detailed history graphs, but only by MAC address. We found the NetXRayPro/100D to be a few revisions behind the 100/S. As a result, its monitoring fell well short of the 100S. The only monitoring that would work, aside from the main gauges, was the history graphs of Ethernet statistics. Cinco Networks claims that its goal is to bring all of its products to the same level of functionality, which is about to increase again by the time you read this, to version 3.0.
Like DominoFastEthernet DA-350 and Shomiti Systems' products, NetXRayPro/100D didn't even blink during our worst-case performance tests. The less-expensive 100S did not fair as well. Its screen froze during some of our more stressful tests, and it displayed some strange utilization numbers, which were off by a few significant digits. Cinco Networks has acknowledged this bug and assured us that it would be fixed promptly. Th
e 100S has the unique ability to capture packets to very large disk files. We did not test this feature, but Cinco claims it can do this at a rate of 30,000 to 45,000 frames per second.
Peter Morrissey is a network systems programmer at Syracuse University. He can be reached at ppmorris@syr.edu.
How We Tested Fast Ethernet Analyzers
We tested our six network analyzers by connecting them to an unmanageable Cisco Systems Fast Ethernet hub and a Netcom Smartbits. We used the Netcom Systems Smartbits Model SMB1000 to generate precise amounts of background and capture traffic. Two of Smartbits' 20 Fast Etherne
t ports were used during testing; one for background traffic and one for capture traffic. We performed two types of tests--a capture test and a utilization test. Both tests were divided into three trials consisting of various levels of background traffic.
For our capture tests, we generated three levels of background traffic. We sent a burst of 3,000 packets, simulating a two-way conversation between two IP addresses, which each analyzer was configured to selectively capture. Trial 1 transmitted 33,500 256-byte packets per second (pps), yielding 75 percent utilization. We thought this test would reflect the traffic that you would find on a very busy Fast Ethernet network.
For Trial 2, we doubled the pps to account for the amount of traffic an analyzer might have to handle on a busy, full-duplex link. It is unlikely that a typical network will approach the 140,000 pps and 94 percent utilization that we generated in Trial 3, but we wanted to see how the analyzers would do in an absolute worst-case scenari
o.
The purpose of the utilization te
st was to check for accuracy. For this test we turned off all filters and transmitted the same three levels of background traffic that we used in the capture test. We then observed the utilization measurements displayed on each analyzer.
REPORTS
Analyize In-Line NAC strategies and products.
ANALYTICS Plan and design your enterprise blade server deployments
InformationWeek U.S. IT Salary Survey 2008
Salaries for business technology professionals are falling. Here's what you need to know in order to make good hiring decisions and personal career choices. Download Today
We tested our six network analyzers by connecting them to an unmanageable Cisco Systems Fast Ethernet hub and a Netcom Smartbits. We used the Netcom Systems Smartbits Model SMB1000 to generate precise amounts of background and capture traffic. Two of Smartbits' 20 Fast Etherne
t ports were used during testing; one for background traffic and one for capture traffic. We performed two types of tests--a capture test and a utilization test. Both tests were divided into three trials consisting of various levels of background traffic.
REPORTS
ANALYTICS
Follow 
