On 7/21/2011 4:25 PM, Martin T wrote:
Chris,
I have no hands-on experience with those servers, but as much as I
have read, they seem to be solid x86 servers. In addition, this "Cisco
Unified Computing System Extended Memory Technology" seems to be a
nice feature.
Pete,
except the "Cisco Unified Computing System Extended Memory
Technology"(at least I haven't seen such feature on other provider
servers), do they have any additional unique features or differences
compared to Dell, IBM or HP servers?
The system management is extremely impressive. Note that the blade
enclosures MUST be networked through the UCS Fabric Interconnect; the FI
is much like a Nexus 5010/5020 but with different code and extra ICs
which provide the UCS management engine. One FI (or one FI redundant
pair) can control multiple UCS blade enclosures (which can each have
multiple servers in them), and it's all centrally managed. I believe
there's a noticeable learning curve at first, but our server peeps seem
to REALLY like it and can turn up new service instances very easily. I
also find the network redundancy to be very well developed; I believe
it's such that the servers don't know there's a redundant network
connected to them (no added complexity of LAG per server).
pt
_______________________________________________
cisco-nsp mailing list [email protected]
https://puck.nether.net/mailman/listinfo/cisco-nsp
archive at http://puck.nether.net/pipermail/cisco-nsp/
