kernelnewbies.kernelnewbies.org archive mirror
 help / color / mirror / Atom feed
From: michi1@michaelblizek.twilightparadox.com (michi1 at michaelblizek.twilightparadox.com)
To: kernelnewbies@lists.kernelnewbies.org
Subject: Kernel latency for handling the Network traffic
Date: Thu, 3 May 2012 21:52:51 +0200	[thread overview]
Message-ID: <20120503195250.GA2189@grml> (raw)
In-Reply-To: <CACYKDtg23F-ssQSE6S1FX0majJYrxmTebPSf0XtbV9L+jxFEsQ@mail.gmail.com>

Hi!

On 09:47 Thu 03 May     , Abu Rasheda wrote:
> On Thu, May 3, 2012 at 6:54 AM,
> <michi1@michaelblizek.twilightparadox.com> wrote:
> > Hi!
> >
> > On 16:33 Thu 03 May ? ? , Suresh Kumar Subramanian wrote:
> >> Hi,
> >>
> >> I am building the router based on linux kernel.
> >>
> >> The hardware details are below,
> >> 2 - 64 bit quad core processor (3Ghz core).
> >> RAM- ?24GB RAM.
> >> PCI express slot- connected with Quad Port 100Mbps Ethernet adapter -2. (so total 8 ethernet interfaces)
> >>
> >>
> >> I just want to calculate the maximum traffic the ?router can handle..?.
> >>
> >> The maximum traffic could be, also 8 ?ports(100Mbps) * 2 directions = 1600Mbps.
> >>
> >> Can this system(kernel + hardware) handle this much traffic. (Assume the best case)?
> >
> > Yes, it can. I have seen a benchmark which basically said that a single quad
> > core cpu with ~3GHz was enough for about 4 links with 10 *gigabit* each.
> 
> What is the packet size ?

It was ~10 million packets per second with 500 bytes packet size, if I
remember correctly. The speed is highly depending on packet size. Actually
packets per second is actually a better unit than (k/m/g)bits per second. I
mostly care about the 500 bytes packet size values in benchmarks because
this is what I think is a good approximate for the average size in most
networks. However, the 64 byte packet size values might also be interesting
when dealing with "weird" applications or DoS attacks.

	-Michi
-- 
programing a layer 3+4 network protocol for mesh networks
see http://michaelblizek.twilightparadox.com

  reply	other threads:[~2012-05-03 19:52 UTC|newest]

Thread overview: 6+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2012-05-03  8:40 Logging sensitive information from kernel (while atomic and uninterrupted) Panagiotis Sakkos
2012-05-03 11:03 ` Kernel latency for handling the Network traffic Suresh Kumar Subramanian
2012-05-03 13:54   ` michi1 at michaelblizek.twilightparadox.com
2012-05-03 16:47     ` Abu Rasheda
2012-05-03 19:52       ` michi1 at michaelblizek.twilightparadox.com [this message]
2012-05-04  2:43 ` Logging sensitive information from kernel (while atomic and uninterrupted) Anirban Roy

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20120503195250.GA2189@grml \
    --to=michi1@michaelblizek.twilightparadox.com \
    --cc=kernelnewbies@lists.kernelnewbies.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).