Most of the (current) things that can be done with software on a NAS
box are close to being released - L2TP, NFAS, etc. In the case of
the VPN stuff, a hardware assist card is needed for good speed.
I believe in a previous post, someone from Livingston said that they
weren't planning on doing multiple IP pools on the PM3 - only 4.
> I think it is more a marketing
> decision - I'm sure RIPv2 could be on the PM-3,
I expect this to happen.
> But it isn't unusual for featurure differentiation to be done with
> different levels of product.
The primary feature differentiation between the PM3 and the PM4 is
the hardware and interconnect density. Anything that requires a hardware
assist makes perfect sense on the PM4 if not the PM3. Something that
is simple software only could be on the PM3. The thought of keeping
certain types of traffic, like NFAS broadcasts, within a chassis is
a happy thing since a congested ethernet wont mean dropped or incomplete
calls. Unfortunately, the price per port sucks.
> Well, this is a HW issue - the PM-4 has the bandwidth to handle these.
I'm not arguing.
> I understand that ComOS 4.1 addresses this - IP Pools are now global and
> centralling managed, so you don't do it per slot.
It doesn't address it until its released and pummeled into QA. :-)
> But they are configured seperately. I think this will converge with time,
> as with the IP Pools.
This would be a very good thing. I don't _care_ if there is a separate
view of Comos running on each blade - but I don't expect my interface
to need to reflect that unless I need very low-level debugging done.
> 1 3T3, 6 14T1, 2 NMC - one slot open, perhaps an Ethernet Switch/Hub blade
> or an OC-3 backhaul.
Although I can agree with Dave Crocker that having a slot burned by a
ds3 card sucks, the lack of spaghetti helps.
If Livingston is clocking all three ds3's off the line0 clock, then
some engineer needs to have a hammer applied to the cranium - repeatedly.
Dropping calls on a t1 is a bad thing. Dropping calls on 2 extra t3's
is nauseating.
> In the long run I'd love to see load-sharing. Especially with higher
> densities, so that you can use both Ethernet interfaces to share the load.
I've had extremely peculiar luck with any form of load sharing - it
makes diagnosis of bad links very hard. A higher density uplink makes
far more sense.
> - -MZ
-- Jeffrey Haas -+- jmh@msen.com -+- http://www.msen.com/~jmh /\/\sen, Inc. "Michigan's Best Run Internet Service Provider." - To unsubscribe, email 'majordomo@livingston.com' with 'unsubscribe portmaster-users' in the body of the message. Searchable list archive: <URL:http://www.livingston.com/Tech/archive/>