About this site
This site is about computer systems performance and benchmarking.
Current Work
I am a Director/Engineer at Nutanix. In my team, we carry out performance engineering on the core Nutanix product, as well as helping customers achieve the best performance from the platform. We believe that a solution can be considered High Performance only if it delivers on these three dimensions.
- Speed – The infrastructure should get out of the way of the underlying hardware and allow the application-software to consume the available hardware performance. IOW the overall solution should be as fast as the available hardware will allow.
- Stability – The system should not exhibit erratic behavior under failure conditions or during typical management operations. Typical failures are disk/host failures. Operations might be migrating workloads from one host to another, taking backups or snapshots, upgrading software etc.
- Scale – The system should be able to consume additional resources without heroic intervention from the operators. IOW If an operator adds an additional 25% capacity – the system should return close to 25% in additional “performance benefit”.
These axis are agnostic in terms of infrastructure deployment and can be applied to classic Enterprise Data Center architecture, Cloud/Web-scale architecture or the emergent combination – currently known as Enterprise Cloud.
Previous Work :
- Performance Engineering at NetApp
- Professional Consulting at NetApp
- Solaris/Oracle Interoperability (Performance and Kernel Panics) at Sun Microsystems
- Or see LinkedIn