Version 6 (modified by Josh Smift, 11 years ago) (diff)


Plastic Slices continuation experiments

This page is the jumping-off point for our continuing work with the ten slices from the Plastic Slices project.

Overview of sites and slices

This is the current mapping of slices to sites to experiments; previous versions of this page have the mapping from previous rounds. More details are also available in the pages below with the details of each round.

Now that we're down to six sites, all sites are in all slices, and BBN only connects via Internet2 (since all of the other sites connect via NLR).

The odd-numbered slices use VLAN 3715, the even-numbered ones use VLAN 3716.

Each site connects as a client (sending data) to each other site in two slices, one on each VLAN. This table shows which slice each site uses to connect to each other site:

to BBN to CLEM to GT to IU to STAN to UTAH
from BBN 101 110 103 102 105 108 107 106 109 104
from CLEM 103 102 105 108 107 106 109 104 101 110
from GT 105 108 107 106 109 104 101 110 103 102
from IU 107 106 109 104 101 110 103 102 105 108
from STAN 101 110 105 108 109 104 103 102 107 106
from UTAH 109 104 103 102 107 106 101 110 105 108

Each slice runs one experiment:

Slice VLAN Experiment
plastic-101 3715 SteadyPing
plastic-102 3716 SteadyPing
plastic-103 3715 SteadyPerf TCP
plastic-104 3716 SteadyPerf TCP
plastic-105 3715 SteadyPerf TCP
plastic-106 3716 SteadyPerf TCP
plastic-107 3715 SteadyPerf UDP
plastic-108 3716 SteadyPerf UDP
plastic-109 3715 SteadyPerf UDP
plastic-110 3716 SteadyPerf UDP

These tables are likely to be folded into the pages for the standing runs, as they'll be changing between runs.

Continuation rounds

These pages contain details of each of the continuation rounds.

First nine rounds

We ran nine rounds of varying duration, to test various things.

Weekly runs

We then started kicking off week-long runs of all ten experiments.

Standing runs

We then switched to having standing runs, of indefinite length, for as long as the composition of the slices remained consistent; as resources were added to (or removed from) the GENI network, we added them.