> I know over the last two years, I've seen several threads on this, > however, I still haven't seen a definitive answer. > > I have a eigerstein box I want to test the throughput, bandwidth, > whatever it's called on. > > What's the best method to simulate a high load? > > Suggestions are welcome!
It depends a lot on the type of load you're trying to simulate. The previously mentioned ttcp is easy to use, but simulates only a single point-point high bandwidth stream. To really test routers & firewalls, you typically need to generate a broad spectrum of traffic, including multiple source/destination IP's, varying packet sizes, varying protocols & source/desination IP's, etc. Typically, the highest load is placed on routers when there are lots of small packets from varying IP's traversing the box...this is much harder to process than the same amount of bandwidth consisting of only large IP packets (there's lots of processing required for each IP packet on a firewall, while simply moving the data around is pretty easy with modern bus-mastering PCI NIC's). There are commercial devices designed to test routers and firewalls, configurable to produce virtually limitless variations of IP traffic. If you can get your hands on one of these, you could test for whatever parameters are important in your application. Besides ttcp, there are some other software based traffic genrators that may be useful...I don't remember any off-hand, but IIRC, there are at least a handful of programs floating around intended to serve as load-generators for web-servers and the like which could probably be pressed into service for testing a router/firewall. If you only use ttcp, make sure you at least play a bit with the packet sizes... Charles Steinkuehler http://lrp.steinkuehler.net http://c0wz.steinkuehler.net (lrp.c0wz.com mirror) P.S. You won't actually need ttcp.lrp unless you want to check bandwidth with the firewall box as an endpoint...while perhaps interesting, this is not exactly "normal" traffic for a firewall/router, so it doesn't necessarily translate directly into useful performance numbers in actual use. ------------------------------------------------------- This sf.net email is sponsored by:ThinkGeek Welcome to geek heaven. http://thinkgeek.com/sf ------------------------------------------------------------------------ leaf-user mailing list: [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/leaf-user SR FAQ: http://leaf-project.org/pub/doc/docmanager/docid_1891.html
