1 | This directory contains the request RSpecs used for the |
---|
2 | three experiments described in EG-EXP-6. |
---|
3 | |
---|
4 | Experiment1/User lnevers: (EG-EXP-6-exp1-*) |
---|
5 | |
---|
6 | - 1 VM w/shared VLAN at BBN |
---|
7 | - 1 VM w/shared VLAN at RENCI |
---|
8 | - SubNet used 10.42.19.0 |
---|
9 | - OpenFlow Controller on port 33017 |
---|
10 | |
---|
11 | Experiment2/User lnevers1: (EG-EXP-6-exp2-*) |
---|
12 | |
---|
13 | - 2 VM w/shared VLAN at BBN (See note 1) |
---|
14 | - 2 VM w/shared VLAN at RENCI |
---|
15 | - SubNet used 10.42.18.0 |
---|
16 | - OpenFlow Controller on port 33018 |
---|
17 | |
---|
18 | Experiment3/User lnevers2: (EG-EXP-6-exp3-*) |
---|
19 | - 1 PG GPO MESO |
---|
20 | - 1 WAPG |
---|
21 | - 1 VM w/shared VLAN at BBN |
---|
22 | - SubNet used 10.42.11.0 |
---|
23 | - OpenFlow Controller is on port 33020 |
---|
24 | |
---|
25 | Note 1: Test case was originally written to include one bare metal node. |
---|
26 | The bare metal nodes are only available via ExoSM, which brings up the |
---|
27 | following options: |
---|
28 | |
---|
29 | a. Replace the BBN SM reservation with a reservation via the ExoSM, |
---|
30 | but this will cause the next step to fail. The RENCI SM reservation |
---|
31 | will report that a sliver by that name already exists, due to the |
---|
32 | ExoSM request. To get around this problem, one option is to replace |
---|
33 | the bare metal node with a VM, so that the requests could be submitted |
---|
34 | as expected to each of the local SMs. |
---|
35 | |
---|
36 | b. Replace the BBN SM reservation with a reservation via the ExoSM, but |
---|
37 | create a new slice and sliver name for the RCI portion. So the BBN |
---|
38 | resources are reserved via the ExoSM, and the RENCI resources are |
---|
39 | reserved via the local RENCI SM. |
---|
40 | |
---|
41 | Option a was chosen for this test. |
---|
42 | |
---|