Find me at VMworld 2007

Next week, I'll be attending VMworld 2007, the virtualization community's annual conference. Actually I won't be the only one given that more than 10,000 people are attending this year! Furthermore, famous people are keynoting:

• Diane Greene, VMware's CEO
• Mendel Rosenblum, VMware's Chief Scientist
• John T. Chambers, Cisco Systems, Inc.
• Patrick Gelsinger, Intel Corporation
• Hector de J. Ruiz, AMD

This year, I'll be giving two talks: "Fast and Easy Disk Workload Characterization on VMware ESX Server" and "ESX Storage Performance - A Scalability Study".

Both talks have more than 500 people registered so that's going to be fun! If you are attending and want to meet up to discuss storage performance issues, you can find me on VMworld Connect.

Here are the detailed abstracts for the talks:

Fast and Easy Disk Workload Characterization on VMware ESX Server

Application performance can be very sensitive to the parameters of the underlying disk subsystem. Therefore, understanding disk I/O workload characteristics for enterprise applications is very important to IT organizations. For example, to optimize RAID stripe size for a particular application requires the knowledge of the size of I/Os generated by the application. Previously, such characterization has been performed either within an operating system using intrusive drivers or outside the system using special hardware.

The session will also demonstrate an efficient implementation of first-level disk I/O workload characterization using online histograms in VMware ESX Server. This allows transparent and online collection of essential workload characteristics for arbitrary, unmodified operating systems running in virtual machines. These capabilities bring disk workload characterization within the grasp of IT administrators thus lowering the bar to application specific tuning.

ESX Storage Performance - A Scalability Study

In this talk, we will present actual performance results of the ESX Server storage subsystem with an emphasis on scalability in distributed environments. The scalability results collected on a 64-host blade cluster with shared storage will be covered for multiple use cases:

1. Distributed storage performance for IO intensive workloads - Covers performance results of various IO block sizes and access patterns with an aim to understand the steady-state scalability, responsiveness and fairness of ESX Server storage.

2. Distributed boot-up performance - Covers boot-up performance of a large number of VMs, helping us understand scalability of massively simultaneous file operations and small IO.

3. Scalable consistency management - Covers performance results of metadata management with advanced distributed locking.

4. Desktop workload performance - Covers performance of VMFS snapshots which enable Scalable Image management, also covering desktop scenarios