|
|
|
|
|
|
|
|
|
|
|
|
| home | help | back | first | fref | pref | prev | next | nref | lref | last | post |
From: Kevin Day <toasty@temphost.dragondata.com> Message-Id: <200107262148.f6QLmZL97633@temphost.dragondata.com> To: lcamtuf@gis.net (Michal Zalewski) Date: Thu, 26 Jul 2001 16:48:34 -0500 (CDT) Cc: stefan@mail.allianztiriac.ro (Stefan Laudat), bugtraq@securityfocus.com In-Reply-To: <Pine.LNX.4.21.0107251732400.747-100000@nimue.bos.bindview.com> from "Michal Zalewski" at Jul 25, 2001 05:38:32 PM MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit > > > The flood is performed via a Gigabit link. The packet rate handling of > > win2k is wonderful, it even beats an OpenBSD 2.8. Kudos to MS guys, > > this one is a real hit. As I couldn't believe my eyes I ran some > > applications on it (crunching queries on the local MS SQL2k server > > etc) and I got timely-fashion responses. > > I believe you are actually testing link layer performance, PCI bus speed > and network cards, not operating systems ;) > Actually, you're probably entering a "livelock" situation. Packets were coming in so fast that the interrupt handler is consuming all your time. Alot of high speed network devices have special modes to prevent this from happening. (Only interrupt when a certain number of packets are in the FIFO, make sure the interrupt isn't asserted for more than x% of the time, etc). This is probably possible on 100MBit links on slow CPUs too. Which network card are you using? I don't ever want to buy one. :) -- Kevin Day toasty@dragondata.com - kevin@stileproject.com
|
|
|
|
|
|
|
|
|
|
|
|
| home | help | back | first | fref | pref | prev | next | nref | lref | last | post |