[Colloquium] Nickolay/MS Presentation/November 21, 2016

[Margaret Jaffey via Colloquium]

This is an announcement of Sam Nickolay’s MS Presentation.

——————————
Department of Computer Science
The University of Chicago

Date:  Monday, November 21, 2016

Time:  9:00 am

Place:  Ryerson 277

Bx/MS Candidate:  Sam Nickolay

MS Paper Title:  Towards Bridging the Gap between Peak and Average Loads on Science Networks

Abstract:
Backbone networks are typically overprovisioned in order to sustain peak
loads. Research and education networks (RENs) are designed to operate at
20-30% loads. For example, Internet2 upgrades the backbone interconnects
when the weekly 95th-percentile load is reliably above 30% of link
capacity. Our analyses of traffic on ESnet between DOE facilities also
show that there is a huge gap between peak and average utilization. As
science data volumes increase exponentially, it is unclear whether this
over provisioning trend will continue in the future. Even if over
provisioning is possible, it may not be the most cost-effective (and
desirable) approach going forward. Under the current mode of free access
to RENs, the traffic at peak load may range from flows that need to be
transferred in near-real time, for example,
for computation and instrument monitoring and steering, to flows that
are less time-critical, for example, archival and storage replication
operations. Thus, peak load does not necessarily indicate the capacity
that is absolutely required. In this paper, we study how data transfers
are impacted when the average network load is increased while the
network capacity is kept at the current levels. We also classify data
transfers into on-demand (ones that are time-critical) and best-effort
(ones that are less time-critical) and study the impact on both classes
for different proportions of both the number of on-demand transfers and
amount of bandwidth allocated for on-demand transfers. We use real
transfer logs from production GridFTP servers for our study. Our results
indicate that the average slowdown experienced by the data transfers is
under 1.5x even when the load is doubled with the network capacity fixed
at the current levels. When the transfers are classified into on-demand
and best-effort, on-demand transfers experience almost no slowdown and
the slowdown experienced by best-effort transfers is under 2x, even when
50% of transfers were treated as on-demand.

Sam’s MS advisor:  Ian Foster

A draft copy of Sam’s MS paper will be available soon.


=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
Margaret Jaffey
Department of Computer Science
Student Affairs Administrator
margaret at cs.uchicago.edu <mailto:margaret at cs.uchicago.edu>
Eckhart 124
773-702-6011
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=



-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.cs.uchicago.edu/pipermail/colloquium/attachments/20161114/fb296b4a/attachment.html>