I'm aware of the CPU difference and I would have thought that 2 modules
sharing 80Mhz would statistically perform better than 4 modules sharing
100Mhz. I'm also aware of the memory capacity difference and again that
statistically, 2 modules sharing 64M wouldn't be much worse than 4 modules
sharing 128M (if your application even requires maxing out the memory).
Can you be more specific on the bus bandwidth difference that you mention?
And, do you mean memory capacity or is there actually a memory bandwidth
value that is important? Any performance stats would be appreciated.
On Sun, 29 Mar 1998, Barry L. Bruins wrote:
> The 3620 has lower performance (bus b/w, memory b/w, etc.) than
> the 3640. So in answer to your first question, it may be a
> penny-wise pound-foolish situation.
> I don't have any suggestions on your second question.
> > 1st question:
> > Since the 3640 chassis ($6500 list) costs more than twice the 3620 chassis
> > ($2900 list), wouldn't it make more sense in many cases to buy 2 3620s
> > instead of 1 3640? (Shouldn't the cost/slot price go down when you
> > increase the number of slots?).
> > That way, it is 'pay as you grow' and you aren't putting all your eggs in
> > 1 basket. I accept the issue of having to manage 2 boxes instead of one. I
> > also accept the issue of traffic between the 2 boxes whereas otherwise it
> > could have stayed within the 1 3640. I am also not considering using these
> > for digital modems which I know requires the 3640.
> > A spare 3620 in case 1 of the other 2 failed would also cost a lot less
> > than a spare 3640.
> > 2nd question:
> > If I want to have a box with 4 ethernet ports and 4 sync (T1) ports, is
> > there any reason not to buy 2 NM-2E2W modules and 4 WIC-1T cards instead
> > of 1 NM-4E and 1 NM-4T?
> > Again, it is cheaper and can be 'pay as you grow' and you don't need near
> > the same value of spares on the shelf. Plus, it is more flexible and the
> > WICs can be cannibalized for other platforms (1600 and 2600).
> > Thanks,
> > Mike.
This archive was generated by hypermail 2b29 : Sun Aug 04 2002 - 04:13:16 EDT