Check out the new USENIX Web site.
FCW '11 Banner Tab
USENIX WIOV '11 Banner

TECHNICAL PROGRAM

All sessions will take place in Salons B–D unless otherwise noted.

Tuesday, June 14, 2011
8:00 a.m.–8:50 a.m.    Morning Coffee and Tea: Served in Salon F and the Ballroom Foyer
8:50 a.m.–9:00 a.m.

Opening Remarks

WIOV '11 Program Co-Chairs: Sanjay Kumar, Himanshu Raj, and Karsten Schwan

9:00 a.m.–10:30 a.m.

I/O Virtualization Architectures

Session Chair: Ada Gavrilovska, Georgia Institute of Technology

SplitX: Split Guest/Hypervisor Execution on Multi-Core
Alex Landau, IBM Research—Haifa; Muli Ben-Yehuda, Technion—Israel Institute of Technology and IBM Research—Haifa; Abel Gordon, IBM Research—Haifa

Read the Abstract | Full paper

View the Slides

Flash Memory Performance on a Highly Scalable IOV System
Peter Kirkpatrick, Adel Alsaadi, Purnachandar Mididuddi, Prakash Chauhan, Afshin Daghi, Daniel Kim, Sang Kim, K.R. Kishore, Paritosh Kulkarni, Michael Lyons, Kiran Malwankar, Hemanth Ravi, Swaminathan Saikumar, Mani Subramanian, Marimuthu Thangaraj, Arvind Vasudev, Vinay Venkataraghavan, Carl Yang, and Wilson Yung, Aprius, Inc.

Read the Abstract | Full paper

VAMOS: Virtualization Aware Middleware
Abel Gordon, IBM Research—Haifa; Muli Ben-Yehuda, Technion—Israel Institute of Technology and IBM Research—Haifa; Dennis Filimonov and Maor Dahan, Technion—Israel Institute of Technology

Read the Abstract | Full paper

View the Slides

10:30 a.m.–11:00 a.m.    Break: Continental Breakfast served in Salon F and the Ballroom Foyer
11:00 a.m.–12:15 p.m.

Invited Talk

Data Center Challenges: Building Networks for Agility
David A. Maltz, Senior Researcher, Microsoft

View the Slides

The key to cost efficiency in data centers is agility—the infrastructure must support allocating resources to services according to their need, and dynamically changing the allocation quickly when needs change. Today, the highest barriers to achieving this agility are limitations imposed by the network, such as bandwidth bottlenecks, subnet layout, and VLAN restrictions. In this talk I'll provide some examples of these barriers and their impact on the services. I'll then present VL2, a practical network architecture that scales to support huge data centers with 100,000 servers while providing uniform high capacity between servers, performance isolation between services, and Ethernet layer-2 semantics, and show how it has helped make data centers run more efficiently.

12:15 p.m.–1:30 p.m.    Lunch: Served in Salons F and I
1:30 p.m.–3:00 p.m.

Performance Management in IOV Systems

Session Chair: Adhyas Avasthi, Nokia Research Labs

Revisiting the Storage Stack in Virtualized NAS Environments
Dean Hildebrand, Anna Povzner, and Renu Tewari, IBM Almaden; Vasily Tarasov, Stony Brook University

Read the Abstract | Full paper

View the Slides

Nested QoS: Providing Flexible Performance in Shared IO Environment
Hui Wang and Peter Varman, Rice University

Read the Abstract | Full paper

View the Slides

Gatekeeper: Supporting Bandwidth Guarantees for Multi-tenant Datacenter Networks
Henrique Rodrigues, Universidade Federal de Minas Gerais (UFMG); Jose Renato Santos and Yoshio Turner, Hewlett-Packard Laboratories (HP Labs); Paolo Soares, Universidade Federal de Minas Gerais (UFMG); Dorgival Guedes, Universidade Federal de Minas Gerais (UFMG) and International Computer Science Institute (ICSI)

Read the Abstract | Full paper

View the Slides

3:00 p.m.–3:30 p.m.    Break: Refreshments served in Salon F and the Ballroom Foyer
3:30 p.m.–5:30 p.m.

Panel/Wild Ideas Session

Panel: Challenges for Virtualized I/O in the Cloud
Participants: Muli Ben-Yehuda, Technion—Israel Institute of Technology and IBM Research—Haifa; Alan Cox, Rice University; Ada Gavrilovska, Georgia Institute of Technology; Satyam Vaghani, VMware; Parveen Patel, Microsoft

View the Slides

As cloud computing becomes more prevalent, I/O virtualization faces stiff challenges to provide performance and quality of service while maintaining the security, reliability, and flexibility promised by system-level virtualization. Further issues arise from the emerging demands for cloud interaction from an increasingly large number of end systems, including smart phones, tablets, and other embedded systems (smart TVs, smart cars, etc.). This panel will discuss the challenges for I/O virtualization for cloud computing infrastructures, posed by demands from compute elements that range from servers, to mobile devices, to embedded systems.

?Need help? Use our Contacts page.

Last changed: 11 July 2011 jp