RE: profiles - a way to simplify iSCSI

["Somesh Gupta" <someshg@yahoo.com>]

Been busy elsewhere a while.

> -----Original Message-----
> From: owner-ips@ece.cmu.edu [mailto:owner-ips@ece.cmu.edu]On Behalf Of
> Matt Wakeley
> Sent: Friday, June 22, 2001 7:14 PM
> To: ips@ece.cmu.edu
> Subject: Re: profiles - a way to simplify iSCSI
> 
> 
> Somesh Gupta wrote:
> 
> > > Mutliple
> > > connections per session is there to take advantage of the 10GBE
> > > link with the shortcomings of TCP,
> > 
> > How? By skipping around the congestion control algorithms? What
> > about the other flows in the network?
> 
> Nobody said anything about skipping around congestion control.
> TCP was not designed for the fast links that are available today, so
> you'll probably need more than 1 connection to attain link rate.
> Whether you have 1 or a thousand connections, congestion control
> will still work (or else it's fundamentally broken anyway).

  Can you please specify what you mean by TCP not being designed
  for fast links (anything other than the congestion avoidance
  alogorithm)

> 
> > > and markers are there to eliminate the cost of memory
> > > subsystems required to buffer out of order TCP frames.
> > 
> > Framing has other benefit but not having a fast memory
> > system cannot be avoided (especially if the inrastructure
> > is to be common.) Also markers and CRCs do not mix well
> > together.
> 
> How do they not mix well?
> 
> > > Eliminate
> > > these, and
> > > you've you'll capitulate to Fibre Channel for anything but 
> slow storage
> > > connects.
> > 
> > I don't think so. The pull of a common infrastructure will
> > be too strong. And in case you have not looked lately,
> > memory is getting fast enough and cheap enough. The GigaHz
> > PCs are driving this.
> 
> Until memory becomes "free", iSCSI implementations will always 
> cost more than
> FC implementations that do not require memory

The question of economics is much wider in scope than whether
you have additional $10 (or whatever) worth of memory. First you look at
the solution cost, of which hardware probably forms less
than a 1/3rd of the cost. Then you go down to specific
components and below. Then you see the economies of scale etc.
How many chips is your R&D cost spread across etc etc.
So to focus on a specific amount of memory at a particular
point in time is not the right way to look at it.

> 
> -Matt

_________________________________________________________
Do You Yahoo!?
Get your free @yahoo.com address at http://mail.yahoo.com