Since I am away this semester and nobody else has had the inclination to carry on the seminars, I'm now going to officially state that this unofficial group has been suspended until such time that someone steps up and wants to take on the (small) work of running it.
In other news, there is now a GSO group specially for CS students. It is officially funded with a substantial amount of capital (taken from student fees). Join the group to help decide what to spend the money on and to partake in the benefits.
Wiki: gsocs.org
Mailing list: www.cs.binghamton.edu/mailman/listinfo/gsocs
Tuesday, October 7, 2008
Tuesday, August 12, 2008
2008-08-15 2PM-3PM, EB G11: Michael Hines on "Post-Copy Based Live Virtual Machine Migration Using Pre-Paging And Dynamic Self-Ballooning"
Michael Hines will be talking about his latest research in a department colloquium. While this isn't an "official" grad seminar talk, everyone is certainly invited to attend. Here's the abstract:
Time: Friday, August 15th, 2008 2:00 PM - 3:00 PM
Location: G-11 Conference Room Engineering Bldg.
Title: "Post-Copy Based Live Virtual Machine Migration Using
Pre-Paging And Dynamic Self-Ballooning"
ABSTRACT
We present the design, implementation, and evaluation of the post-copy
based method for the live migration of virtual machines across a
Gigabit LAN. Live migration is a mandatory feature of modern
hypervisors today. It facilitates server consolidation, system
maintenance, and lower power consumption. Post-copy refers to the
deferral of the ``copy" phase of live migration until the virtual
machine's CPU state
has already been migrated. This enables the migration daemon to try
different methods by which to perform the memory copy. With
post-copy, we seek the goal of a ``win-win'' by deterministically
guaranteeing at most the same migration time of the static, pure
stop-and-copy method. Post-copy also provides the downtime and
liveness benefits of pre-copy, without relying on an unbounded
iterative process. We facilitate the use of post-copy with a specific
instance of adaptive pre-paging (also known as adaptive remote
paging). Pre-paging is capable of eliminating all duplicate page
transmissions and removing any residual dependencies. Our algorithm is
able to minimize the number of page faults to less than 20% of the
Virtual Machine's writable working set. Finally, we facilitate both
the original pre-copy and post-copy schemes with the use of dynamic,
periodic self-ballooning. This prevents the migration daemon from
transmitting unnecessary free pages in the guest system. This also
noticeably speeds up both migration schemes with very negligible CPU
degradation to the processes running within the Virtual Machine. We
implement post-copy on top of the Xen Hypervisor and benchmark its
behavior against the existing pre-copy method.
For more information please visit:
http://osnet.cs.binghamton.edu/publications/TR-20080702.pdf
Time: Friday, August 15th, 2008 2:00 PM - 3:00 PM
Location: G-11 Conference Room Engineering Bldg.
Title: "Post-Copy Based Live Virtual Machine Migration Using
Pre-Paging And Dynamic Self-Ballooning"
ABSTRACT
We present the design, implementation, and evaluation of the post-copy
based method for the live migration of virtual machines across a
Gigabit LAN. Live migration is a mandatory feature of modern
hypervisors today. It facilitates server consolidation, system
maintenance, and lower power consumption. Post-copy refers to the
deferral of the ``copy" phase of live migration until the virtual
machine's CPU state
has already been migrated. This enables the migration daemon to try
different methods by which to perform the memory copy. With
post-copy, we seek the goal of a ``win-win'' by deterministically
guaranteeing at most the same migration time of the static, pure
stop-and-copy method. Post-copy also provides the downtime and
liveness benefits of pre-copy, without relying on an unbounded
iterative process. We facilitate the use of post-copy with a specific
instance of adaptive pre-paging (also known as adaptive remote
paging). Pre-paging is capable of eliminating all duplicate page
transmissions and removing any residual dependencies. Our algorithm is
able to minimize the number of page faults to less than 20% of the
Virtual Machine's writable working set. Finally, we facilitate both
the original pre-copy and post-copy schemes with the use of dynamic,
periodic self-ballooning. This prevents the migration daemon from
transmitting unnecessary free pages in the guest system. This also
noticeably speeds up both migration schemes with very negligible CPU
degradation to the processes running within the Virtual Machine. We
implement post-copy on top of the Xen Hypervisor and benchmark its
behavior against the existing pre-copy method.
For more information please visit:
http://osnet.cs.binghamton.edu/publications/TR-20080702.pdf
Tuesday, July 8, 2008
2008-07-09 6:00PM-7:30PM, EB Q3: Calendaring Roundtable
I've been thinking about calendaring lately and we could have
another "roundtable" style discussion about calendaring in the CS
graduate student environment. If everyone's already allocated this
timeslot (2008-07-09 @6PM), maybe we can go ahead with a meeting about
using calendar tools.
I personally use Google calendar and hook it into my desktop calendar
software. I'd be interested in hearing about other solutions folks use
(from software-based solutions to paper-based to solutions to
neurologic-based solutions).
another "roundtable" style discussion about calendaring in the CS
graduate student environment. If everyone's already allocated this
timeslot (2008-07-09 @6PM), maybe we can go ahead with a meeting about
using calendar tools.
I personally use Google calendar and hook it into my desktop calendar
software. I'd be interested in hearing about other solutions folks use
(from software-based solutions to paper-based to solutions to
neurologic-based solutions).
Thursday, June 26, 2008
To be rescheduled: Avadh Patel on GTK+ and Cairo (room TBD)
Avadh Patel will speak next about the GTK+ and Cairo libraries for
building cross-platform graphical applications with high quality 2d
graphics.
This talk is being rescheduling...
building cross-platform graphical applications with high quality 2d
graphics.
This talk is being rescheduling...
Thursday, March 20, 2008
2008-03-20: Semantic Web (unofficial grad seminar talk)
The BUCS grad seminar folks are invited to attend a talk in G11 at 6:30PM today, Thursday, March 20, 2008.
The talk will be on the topic of the Semantic Web.
The talk will be on the topic of the Semantic Web.
Tuesday, March 18, 2008
Next Talk: Mikhail Gofman on RBAC on 2008-03-19 at 6:30PM in EB G11
The next talk will be held (tomorrow) Wednesday, March 19, 2008 in the large conference room (EB G11 in the Engineering Building on the Binghamton University main campus).
Mikhail will talk about the research he's done for his MS, which he will be defending on Thursday (March 19).
The title of his talk is EFFICIENT POLICY ANALYSIS FOR ADMINISTRATIVE ROLE BASED ACCESS CONTROL.
Pizza and soda are planned to be made available. Donations (of a few dollars) are requested, but not required, for the consumption of food.
Mikhail will talk about the research he's done for his MS, which he will be defending on Thursday (March 19).
The title of his talk is EFFICIENT POLICY ANALYSIS FOR ADMINISTRATIVE ROLE BASED ACCESS CONTROL.
Pizza and soda are planned to be made available. Donations (of a few dollars) are requested, but not required, for the consumption of food.
Monday, February 18, 2008
2008-02-21 (5PM-6PM), EB G11: Computing with Light
On Thursday, February 21, 2008, we will have the next BUCS talk. The talk will cover "Computing with Light" and show how one can solve NP-complete problems in polynomial time (and exponential space) using a photo copier.
Monday, January 28, 2008
2008-02-14 (4:15-5:50): LaTeX panel discussion
On Thursday, Feb 14 in the big conference room (G11) in the CS lab space, I'll be hosting a panel discussion about LaTeX. Cenk, Michael Hines, Mikhail Gofman, and/or I will talk a little about LaTeX and answer questions from the audience.
UPDATE: changed the date after finding out the PHD exam was being held on the 7th.
UPDATE: changed the date after finding out the PHD exam was being held on the 7th.
Wednesday, December 5, 2007
Jason Moore and Lee Seversky on Java/OpenGL: EB Q3 at noon on Wednesday, 2007-12-05
Presenters:
Jason Moore, Air Force Research Laboratory
Lee Seversky, Air Force Research Laboratory
IN EB Q3 NOON today...
Jason Moore and Lee Seversky will be presenting an overview of the research in visualization and 2D/3D graphics being performed at the Air Force Research Laboratory in Rome, NY. The talk will include demonstrations of applications written entirely in JAVA using OpenGL which also run in real-time!
Jason Moore, Air Force Research Laboratory
Lee Seversky, Air Force Research Laboratory
IN EB Q3 NOON today...
Jason Moore and Lee Seversky will be presenting an overview of the research in visualization and 2D/3D graphics being performed at the Air Force Research Laboratory in Rome, NY. The talk will include demonstrations of applications written entirely in JAVA using OpenGL which also run in real-time!
Tuesday, November 27, 2007
Cenk's defense: 2007-11-28 @3:30PM in EBG11
Değer Cenk ERDİL
erdil@cs.binghamton.edu
Department of Computer Science
Watson School of Engineering and Applied Sciences
Binghamton University
Date: Wednesday, November 28th, 2007 Time: 3:30pm Location: Engineering Building G-pod Conference Room
ABSTRACT
As grids scale to include more individual resources, improved grid middleware services that eliminate the need for structure and centralization becomes more important. Moreover, the most effective grid middleware services will be adaptive, reacting to their highly dynamic environments.
One important service that must be distributed and scalable is grid resource scheduling. In hybrid grids, resources and jobs are likely to be non-uniformly distributed in space and in time, and thus no single approach to tracking resource information will be effective in all places at all times. Non-uniform information dissemination helps enable large-scale and dynamic grid resource scheduling, with less packet overhead and better localized coverage, which is especially useful when the grid structure is mostly unknown, such as in hybrid grids.
This dissertation describes improved information dissemination protocols that adapt to their environment using feedback from the system. Thus, grid nodes produce different dissemination protocols on-the-fly, where each protocol individually reflects both the characteristics of particular resource and load distributions, and the policies of autonomous grid nodes or regions. The adaptive information dissemination protocols can result in much less packet overhead, with comparable query satisfaction rates, compared to best-case non-adaptive protocols that may be configured for each particular grid resource and load scenario.
Vinay's dissertation defense: 2007-11-28 @8:30AM in EBG11
Routing and Traffic Engineering in Multi-hop Wireless Networks: An optimization based approach
Vinay Kolar
vinkolar@cs.binghamton.edu
Dept. of Computer Science, SUNY, Binghamton.
Advisor: Dr. Nael Abu-Ghazaleh
Date: Nov 28, 2007, Wednesday Time: 8:30 A.M. Room: EB-G11
Abstract
Multi-hop wireless networks (MHWNs) attract significant interest due to the minimal infrastructure demands and their potential in supporting mobile and pervasive computing. The high demand placed by a growing user base on the limited available bandwidth places a premium on effective communication and networking for MHWNs. The dissertation targets developing a formally grounded approach to solving routing problems in MHWNs, while taking into account the effects of interference. The work builds on recent efforts in the networking community to express a network as an optimization problem, and decomposing the formulation to provide distributed protocols. A successful model can then be applied to: (1) analyze the performance and capacity of existing protocols; (2) develop protocols for traffic engineering and admission for static networks; and (3) develop formally grounded and near-optimal distributed routing protocols.
In MHWNs, the problem is substantially more complicated than the wired problem because of interference. Interference is exhibited at many levels, leading to effects such as uncontrolled contention and unfairness. The approach taken by the dissertation is to break the problem into multiple layers.
Firstly, the dissertation proposes a Multi-commodity flow based routing model that produces interference-separated routes. Interactions between multiple routes are analyzed and effective objective function design is proposed that maximizes the throughputand minimizes the end-to-end delay.
Nevertheless, such a model is directly employable only in smaller networks due to its NP-hard nature. The second contribution of the dissertation is to approximate the routing model to a polynomial time algorithm. A decomposition based approach is followed to formulate a low-complexity model by applying domain-specific heuristics.
The simplistic scheduler assumptions in the routing model limits its accuracy in practice. A low-complexity scheduling model is proposed to capture the key scheduling interactions in CSMA based schedulers, like IEEE 802.11, and this model is integrated with the routing model.
An approach to improve the accuracy of the scheduling model, while preserving a low run-time, is presented. “Interaction graphs” are proposed to capture the scheduling characteristics of CSMA based schedulers. Fairness in CSMA based scheduler for contention of the wireless channel is modeled using Renewal Theory and Continuous-time Markov Chain. Finally, throughput estimation models are proposed for various categories of interactions that have been identified in MHWNs.
Saturday, November 24, 2007
Wednesday 11/28/07 talks
Both Vinay and Cenk have dissertation defenses on Wednesday. In lieu of a regular talk, we'll attend these defenses and support our fellow students!
Vinay will go on 8:30am, Cenk's is at 3:30pm.
I don't know the room assignments, but there should be invitations posted around the engineering building on campus.
Vinay will go on 8:30am, Cenk's is at 3:30pm.
I don't know the room assignments, but there should be invitations posted around the engineering building on campus.
Thursday, November 1, 2007
Video posted of power circuit talk
I have posted the first video. It's from the 2007-10-31 talk by Michael Hines about Controlling Power with the Computer.
The video didn't come out as great as I would have liked. In the future I'll do better sound checks and hopefully will find a better recording tool that make just create avi/divx files.
The video didn't come out as great as I would have liked. In the future I'll do better sound checks and hopefully will find a better recording tool that make just create avi/divx files.
Tuesday, October 30, 2007
Wed, Oct 31, Noon: Switching Power with the Computer
Michael Hines (a doctoral student in the CS department) will be talking tomorrow (Wed, Oct 31) about a hands on project about using a computer to control a power switch. It should be a fun talk and will be held in the new CS department research area (the old South Pod).
Notes on the talk are online: http://hinespot.net/circuit.php
Notes on the talk are online: http://hinespot.net/circuit.php
Friday, October 19, 2007
Skipping a week
With the Ph.D. Qualification exams, it's probably best to skip the talk that would have been made on Oct 24.
Therefore, the next talk will be on Oct. 31 at noon, and it should be held in the conference room in the new CS research lab area (the G section of the Engineering Building)
Therefore, the next talk will be on Oct. 31 at noon, and it should be held in the conference room in the new CS research lab area (the G section of the Engineering Building)
Wednesday, October 17, 2007
CVS/BZR slides posted
The slides of the CVS/SVN/BZR talk have been posted.
Thursday, October 11, 2007
Talk for Wednesday, October 17 -- CVS vs. BZR
Michael Head will be talking about modern and historical revision control systems.
The talk will be held in EB N25 at noon on October 17, 2007.
The talk will be held in EB N25 at noon on October 17, 2007.
Wednesday, October 10, 2007
Friday, October 5, 2007
Next Two Talks: Wednesdays Oct 10 and 17, Noon N25
The next two talks have been scheduled from Noon to 1PM in EB N25. Cenk will talk on the 10th, and the talk on the 17th is TBD.
Monday, October 1, 2007
Seminar tonight -- Brent Rood on Characterizing Grid Resources
Brent will speak tonight (2007-10-01) from 5-6PM in EB N25. Pizza and soda has been ordered from Nirchi's.
Brent's talk is about his current research that was presented last week at the GRID 2007 conference.
Brent's talk is about his current research that was presented last week at the GRID 2007 conference.
Subscribe to:
Posts (Atom)