Monthly Archives: September 2008

Xen Summit North America 2009: Hold the Date

For those of you with calendars reaching into 2009, please reserve February 24 – 25 for the first Xen Summit 0f 2009 hosted by Oracle in Redwood Shores, CA. We are actively putting the schedule together for this event and will be calling for volunteers to work on the Program Committee later this month. Please make sure to update your calendar.

Also, Xen Summit Tokyo is currently open for registration at https://www.regonline.com/xsasia08 and we continue looking for speakers for this event.

New Web Site is LIVE!

The new website is live; please send me any problems you find so I can make fixes.

Current Issues:

1) The Wiki did not transfer over to the new look & feel – not sure why but looking into it

2) The Mailing Lists registration page for a new user has issues ; file access is the problem here but chagne is easy once I can get to the proper files

Project Snowflock Available

Project Snowflock is now available to the general public. We’re making a binary and source release, under the GNU General Public License (GPL). The release is available at http://compbio.cs.toronto.edu/snowflock.

Briefly, Snowflock lets you clone Xen VMs into dozens of identical replicas running in different hosts. Snowflock can do this in less than a second and with very low runtime overhead. With Snowflock you can, for example, perform parallel computations on the fly by scaling “instantaneously” your computing footprint in a shared cluster.

Snowflock is a research prototype, hence the 0.1 major-minor. A minimum degree of experience with Xen and Linux is necessary to use the system. The contact address for snowflock is snowflock-users@cs.toronto.edu

More technically:
Snowflock is our prototype implementation of the / Impromptu Cluster (IC)/ abstraction. In an IC, an application encapsulated inside a virtual machine (VM) is swiftly forked into multiple copies that execute on different physical hosts, and then disappear when the computation ends. ICs simplify the development of parallel applications and reduces management burden by enabling the instantiation of new stateful computing elements: workers that need no setup time because they have a memory of the application state achieved up to the point of forking. This approach combines the benefits of cluster-based parallelism with those of running inside a VM.

Snowflock provides swift parallel VM cloning that makes it possible for Internet applications to deliver near-interactive performance for resource-intensive highly-parallelizable tasks. Snowflock makes use of four key techniques: /VM descriptors/ (condensed VM images that allow for sub-second suspension of a running VM and resumption of a of replicas); a /memory-on-demand/ subsystem that lazily populates the VM’s memory image during runtime; a set of / avoidance heuristics/ that minimize the amount of VM memory state to be fetched on demand; and a /multicast distribution/ system for commodity Ethernet networking hardware that makes the overhead of instantiating multiple VMs similar to that of instantiating a single one.

Use, enjoy, give us some feedback, and contribute to mankind :) The Snowflock team

USENIX LISA 08 Conference

The Xen.org community will again be running a full day training session at an upcoming USENIX Event – LISA ’08. For those of you going to event or looking for a hands on approach to Xen training, the community will be hosting the following training session on Monday November 10th from 9am to 5pm – Introduction to the Open Source Xen Hypervisor. The class is be taught by myself along with 2 co-authors of the Running Xen book out of Clarkson University: Zach Shepard and Wenjin Hu.

Here is the conference information on LISA:

———————————————————–
SAVE THE DATE
LISA ’08:
22nd Large Installation System Administration Conference http://www.usenix.org/lisa08/progm
November 9-14, 2008, San Diego, CA
Early Bird Registration Deadline: October 17, 2008 Sponsored by USENIX and SAGE
———————————————————–

LISA comes to San Diego,  November 9-14, 2008.  As always, the breadth and quality of this year’s tutorials, refereed papers, invited talks, and participants are excellent. Some highlights:

Take advantage of the popular 6 days of training. Select from over 50 tutorials taught by highly expert instructors, including:

— Mark Burgess on Integrating Cfengine into Organizational Service Management
— Tom Christiansen on Advanced Perl

New in 2008: Virtualization and Solaris Tracks

Get 6 days of focused instruction by industry leaders including:

— Richard McDougall on VMware ESX Performance and Tuning
— James Mauro on Solaris 10 Performance, Observability, and Debugging

— And more . . .

The full training program can be found at http://www.usenix.org/lisa08/training/

In addition to the training, 3 days of technical sessions include top-notch refereed papers, informative invited talks, expert Guru Is In sessions, and a poster session: http://www.usenix.org/events/lisa08/tech/

* Our 20+ invited talks feature our most impressive slate of speakers to date. They include:

— Keynote Address: “Implementing Intellipedia Within a ‘Need to Know’
Culture,” by Sean Dennehy, Chief of Intellipedia Development, Directorate of Intelligence, U.S. Central Intelligence Agency

— Plenary Session: “Reconceptualizing Security,” by Bruce Schneier, Chief Security Technology Officer, BT

–Plenary Session: “The State of Electronic Voting, 2008,” by David Wagner, University of California, Berkeley

LISA is the premier forum for presenting new research in system administration. We selected papers  showcasing state-of-the-art work on topics including configuration management, parallel systems deployment, virtualization, and security.

Explore the latest commercial innovations at the Vendor Exhibition:
http://www.usenix.org/events/lisa08/exhibition.html

For complete program information and to register, see http://www.usenix.org/lisa08/progm

Early registration discounts for LISA ’08, taking place November 9-14, 2008, in Dallas, are now available. Register by Friday, October 17, and save up to $300!

—————————————————————————
WHAT:    LISA ’08: 22nd Large Installation System Administration
Conference
WHEN:    November 9-14, 2008
WHERE:  San Diego, CA
WHO:       System Administrators, Network Administrators, CIOs, CTOs,
Researchers, Tool Providers, Support and Help Desk personnel, etc.
WHY:      To get to and stay on the cutting edge of computer system
administration
HOW:      http://www.usenix.org/lisa08/progm

Xen.org Website Update – Final Review

For those of you who have not looked over the changes to the www.xen.org website, please take a look as I plan to switch to the new site this weekend or early next week. If you have any comments, please let me know.

Test Site: http://staging.xen.org

For more details on this project:  http://blog.xenproject.org/index.php/2008/08/26/preview-the-new-xenorg-website/

Xen 3.3 Feature – C & P State Power Management

One of the notable features of Xen 3.3 is to incorporate full support for processor power management features, C-states and P-states. Power management is getting more crucial not only for clients, but also for servers. C/P is from ACPI nomenclature which stand for different set of power/thermal technologies. Combining both, Xen 3.3 is now expected to achieve far better performance per watt.

The Xen 3.3 idle governor can support full line of C-states with great power reduction, especially for idle power consumption. The idle governor is triggered when the CPU is fully idle, and then the governor chooses the appropriate low power state based on the power budget and latency tolerance accordingly. The deeper C-state is, less power is consumed with longer entry/exit latency. The previous Xen releases only supported C1, as triggered by the HLT instruction.

The Xen 3.3 on-demand governor monitors CPU utilization in small intervals and dynamically chooses a suitable operation point with lower power and negligible performance impacts. P-states are set of operation points about CPU frequency/voltage, which can help reach optimal performance at the lowest power even when the system is in load line. Unlike C-states, P-state transitions can be triggered at fine-grained level, upon fast-changing workload characteristics. CPU vendors provide smart hardware level enhancements  to ensure fast on-demand frequency/voltage change, like Intel’s Enhanced Intel SpeedStep@ and AMD’s PowerNow!@.