North American Network Operators Group

Date Prev | Date Next | Date Index | Thread Index | Author Index | Historical

Re: Jumbo Frames (was Re: MAE-EAST Moving? from Tysons corner toreston VA. )

  • From: Stephen Sprunk
  • Date: Mon Jun 19 13:40:53 2000

Sez "Roeland Meyer (E-mail)" <[email protected]>
> Actually, my testing shows a falure to utilize even 100baseTX
> fully. Even in a switched FDX environment (no collisions) I can't
> achieve line rate without bumping the packet size up. Considering
> that the smallest box is a quad-CPU SMP machine (550Mhz), I don't
> think that there is a CPU shortage <grin>.

It's a rare event to see any server get line rate on any media.

If you bump up the media speed, you'll generally see more throughput.  A
box which can't saturate a FE link can often manage >100mbit/s on a GE
link.  The equivalent was true way back when FE was new.

> I don't have a 10gig-E system, but I wonder about going there
> when I can't even get gig-E to work efficiently.

Perhaps you should talk to your server/app vendors about that.

> If vendors want to sell 10gig-E they should be concerned about
> exactly this point.

The point is exactly the same point that FE and GE had.  Initially TGE
will be used as an aggregation point for lots of lower-speed user/server
ports.  As servers speed up to match the network's capabilities, TGE
will migrate out to the servers and the network core will progress to an
even faster technology (Hundred Gig E?).

> Joe SOHO isn't going to buy it anyway.

Joe SOHO is generally still using 10mb hubs and isn't relevant here.

> Joe Enterprise isn't going to spend the extra money unless he can
> see some real benefit,

That's funny, the biggest cry for TGE I hear is coming from Joe
Enterprise, who is complaining that a GE network isn't fast enough to
push his hundred-TByte backups and GE-connected workstations.

> and Joe dot-com ain't going to do it unless it is measurably faster
> than gig-E (which it won't be with MTU=1500).

It will be 10x as fast, regardless of the MTU.  If Joe's servers can't
keep up, that doesn't change how fast TGE runs.

> I can aggrigate 3-5 gig-E links to get the same troughput, by
> adjusting MTU, and not pay the 10gig-E meal-ticket.

If you want to go 10x as fast as GE, you will need either (a) 10 GE
links with perfect loadsharing or (b) TGE.  If you only care to go
5Gbit/s, that's not "the same throughput" simply because your servers
can't keep up with TGE.

> BTW, the selling feature on gig-E is link aggrigation, built into the
> (over Fast-E), there is no similar feature enhancement for
> 10gig-E, AFAICT.

To what link aggregation feature do you refer?  I'm not aware of any
functional difference between the aggregation capabilities of GE and FE.

> Evenso, it is still limited by MTU size.

Hindered, not limited.  There are devices that can fill a GE link with
64byte frames; hardware to do the same will undoubtedly appear for TGE.

A 1500byte MTU is definitely a problem for server vendors, but claiming
that faster media is pointless because of it is hardly realistic.
Perhaps those vendors will get involved in the IEEE TGE process and get
jumbo frames standardized.


     |          |         Stephen Sprunk, K5SSS, CCIE #3723
    :|:        :|:        Network Design Consultant, HCOE
   :|||:      :|||:       14875 Landmark Blvd #400; Dallas, TX
.:|||||||:..:|||||||:.    Email: [email protected]