Feb 7, 2019 8:08:21 AM by Gary Jerep
Real stories. Real people. Real situations. With Toad solutions.
Here’s a recent Toad experience from less than a month ago, at the time of this posting. I am speaking to the Director of a database administration team for a financials company based in the western U.S. We’ll call him Don. Don and his team members already own seats of our Toad for Oracle DBA Edition. But instead of the usual request for us to conduct a “tips and tricks” session, Don wants to focus on Benchmark Factory, one of the components of their Toad edition.
Don’s DBA team uses Benchmark Factory regularly to run activity loads on their databases before and after proposed changes. The team can check for any adverse impact to database performance before any changes get pushed officially to production. Don already knows the product very well.
So today, Don explains what he wants out of our consultation: He is buying additional licenses of Benchmark Factory to help with a database migration project. The company is taking one of their in-house-developed application stacks and migrating its current database back-end from Oracle to Postgres. Don wants to make sure that Benchmark Factory will work on the Postgres side the same way that his team uses it for Oracle. Additionally, they want to ensure that performance on Postgres compares favorably with that on their Oracle back-end, and to vet out any unforeseen “gotchas” before migration. The short demonstration that follows satisfies the team that Benchmark Factory will indeed do what they need, including a comparison of the performance differential between the two database platforms.
After the demo, Don opens up about some of the other projects that their IT department is pursuing, including a recently completed hardware migration. Last year, top management mandated that the company pull the plug on their legacy hardware iron and port over to distributed x86-based servers. They completed that migration late last year.
“We saved hundreds of thousands of dollars…twice.” Don says matter-of-factly.
Now, saving big bucks on a platform migration isn’t news to my ears. Happens all the time. But what intrigues me right now is the one little word Don mentioned at the end of his sentence.
Don explains that the cost differential in going from their legacy servers to x86 boxes was close to a million dollars. That was savings number one. But Benchmark Factory gave them savings number two, an additional six-figure dollar amount. Don says all this with remarkable coolness, which contrasts sharply with my party popper reaction that follows.
“OK, whoa. BMF did WHAT?!” I shouldn’t be surprised, but I am. ANY product that saves you ten thousand times its purchase price would give anyone an adrenalin rush.
Don continues in a nonchalant tone that suggests he saves his company money like this every day. Yawn! The hardware vendor gave Don’s company a recommendation of how many x86-based servers are needed to equal the computing power of the legacy database server environment. Not that Don’s company didn’t trust their hardware vendor, please note. But Don wanted to do his own due diligence. He did not blindly accept the vendor’s numbers. Don’s DBA team came up with their own calculations, and identified several possible scenarios with varying numbers of x86 servers. Each of these scenarios involved a number of servers that was below the vendor-recommended number. Don and team defined virtual database server environments to test each one’s computing power. One environment matched the vendor’s original proposed configuration. A few more were created to test each of the team’s own scenarios.
Don used Benchmark Factory to run industry-standard activity on the old legacy system to benchmark current database performance and throughput under load. Then they executed the same database activity load on each of the virtual environments. Benchmark Factory reports were showing that database backend performance and throughput within each of Don’s scenarios were comparable (to varying degrees) to the database throughput they saw on their old hardware system. This helped Don and team narrow in on the actual number of new servers to purchase.
Ah, but these are virtual environments with virtual results. We should probably test on the real thing, right? So Don worked with their hardware vendor to install their proposed system, and did one last round of Benchmark Factory tests to confirm their conclusions on the real physical environment before purchasing.
The bottom line? Don was able to stand up an x86-based database server environment whose database performance matched that of his old iron, at a savings of more than $100,000 below the vendor’s original estimate. Wow.
Benchmark Factory. Six figures. What’s in your wallet?
Tags: Benchmark Factory
Written by Gary Jerep
As a Software Consultant within Quest, Gary Jerep has over 20 years of experience assisting DBAs, Developers and Analysts with relational database needs. Prior experience includes systems design and analysis, operations research, Business Intelligence, and end-user experience monitoring.