Thursday, June 1, 2017

Monster Oracle Database Virtual Machines

I recently finished a whitepaper on the performance of really big Oracle database virtual machines. I highly recommend that you take a look at the paper for the full details, but a quick summary is that performance of large database virtual machines is good.  This round of testing was using vSphere 6.5 on four generations of four-socket Intel based servers.  The biggest size virtual machine that could be run without using any of the hyperthreads was tested on each host and compared:

As the power of servers has increased, the performance of large virtual machines has also increased.

Another way of looking at the performance increases is to keep the size of the virtual machine the same, and move it from older to newer hardware.  Specifically if you compare the performance of the same 40vCPU VM on Westmere-EX vs Broadwell-EX you find an increase in performance of 42%:

The older Westmere-EX based server had 10 cores per socket, so the 40 vCPU VM was using all the cores.  The newer Broadwell-EX based server has 24 cores per socket, so the 40 vCPU VM fits easily.  So in addition to the 40 vCPU VM performing 42% better, there is also room for more VMs on the Broadwell-EX server.  CPU utilization of the Broadwell-EX server was just 42% when the 40 vCPU VM was fully utilized because the host has 96 cores.

Thursday, February 19, 2015

So Say SMEs Cover CMDs/Sec

Our performance counter of the week is a key storage related metric - CMDs/Sec.  This is part of our new effort to cover a a new performance counter every week - along with a little bit of other virtualization and cloud stuff as needed.  We secretly hope to make you all esxtop performance experts.

Friday, February 13, 2015

So Say SMEs and CoStop Performance Counter

This week on So Say SMEs, Kong and I discuss what the costop or %costop performance counter means and how you can use it to resolve performance issues on your vSphere / ESXi environment.  You can view costop in esxtop on the main CPU screen that comes up by default.

Thursday, February 5, 2015

Intro to CPU Ready Time in ESXTOP

This week in our So Say SMEs episode we have a quick discussion about the CPU Ready time performance counter in esxtop.  This has been covered by some other great blogs and esxtop performance guides.  Here we have a quick intro conversation and also talk about the Super Bowl - because we have to talk about the Super Bowl.

Tuesday, January 27, 2015

So Say SMEs Returns for Third Season

Kong and I have returned for a third season of So Say SMEs.  We have a few changes for season 3, but we will continue to talk cloud and virtualization with some geek and sports talk too.  In terms of changes - Kong has moved over to SolarWinds and we going to focus a bit more on performance when we get the chance.

 My voice is still not 100% in this video, but that is a story for another day!

Tuesday, September 17, 2013

Flash Cache at Chuy's

As Kong continues his job search we found a place that was hiring for this week's episode of  So Say SMEs. Chuy's is of course the perfect place to discuss the new vSphere Flash Read Cache feature aka vFRC.

I decided to wear a vintage work shirt this week - from a company that no longer exists.  Anybody remember Altiris?

Friday, September 6, 2013

Let The TPC-VMS Benchmark Wars Begin

In the benchmark world, the Transaction Processing Performance Council (aka TPC) has long been the owners of the premier performance benchmark.  I'm referring to the TPC-C database performance benchmark where millions have been spent to prove performance excellence.

A new benchmark from TPC that is designed to focus on the performance of virtualized databases was created called TPC-VMS.  And very recently the first TPC-VMS result was published by HP using vSphere 5.5.

This means that the HP and VMware vSphere solution has the "Leadership" position for the TPC-VMS benchmark - as of right now.

What happens next will be interesting.

The first published result establishes the target for everybody else.  Will there be a bunch of publications over the coming months?  Each a little bit better.  Or will it take some time for anybody to be able to surpass this initial publication?

It is also of course possible that nobody else will ever publish a TPC-VMS result.  Sometimes new benchmarks just never get established.  If this is the case, then vSphere would hold the number one position forever.  But this isn't what we want.

Ideally the healthy competition of a benchmark fosters innovation, software and hardware improve, costs come down, and customers benefit.