... | ... | @@ -15,17 +15,15 @@ The Utah Emulab cluster has 160 PC nodes (**pc401-pc560**) consisting of: |
|
|
|
|
|
All interfaces on these machines are connected to gigabit ports on
|
|
|
[ProCurve 5412zl</span>](http://www.procurve.com/products/switches/HP_ProCurve_Switch_5400zl_Series/overview.htm)
|
|
|
switches (specifically, J8702A modules). There are three new
|
|
|
switches (specifically, J8702A modules). There are three
|
|
|
experimental-net switches (**procurve3-procurve5**) for these nodes, each
|
|
|
with approximately 240 ports. A fourth new switch (**procurve1**) acts
|
|
|
with approximately 240 ports. A fourth switch (**procurve1**) acts
|
|
|
as the center of a 'hub and spoke' topology for the experimental network:
|
|
|
it is connected to the new switches and to the existing gigabit switch
|
|
|
(**cisco8**) at 24.4 Gbps each, and to the existing 100Mbps switches using
|
|
|
4Gbps to 8Gbps trunks.
|
|
|
it is connected to the other three ProCurve switches at 24.4 Gbps each (via a 4 x 10Gb module, but the backplane can only manage 24.4 Gbps aggregate) and to the rest of the testbed via 4 Gbps and 8 Gbps trunks.
|
|
|
|
|
|
Most nodes have 4 experimental-net interfaces, but some have 5.
|
|
|
The d710 nodes have their interfaces 'striped' across the switches:
|
|
|
each PC has at least one interface on each of the three
|
|
|
each PC has at least one interface on each of the three hub
|
|
|
switches, so that large LANs can be constructed of all nodes, without
|
|
|
needing to use any inter-switch bandwidth. The layout is:
|
|
|
|
... | ... | |