When Vertica was acquired by Hewlett Packard, we were told that we needed to change our messaging about commodity hardware. HP didn’t feel comfortable with the phrase, because it implied that underlying infrastructure was not a key differentiator.
But as a true software company who designed every line of code to be optimized for infrastructure efficiency, Vertica was determined to ensure that, no matter what choice our customers made for servers and storage, Vertica would deliver the best possible performance. We had no incentive to push for more compute capacity because we didn’t make any money on the underlying infrastructure. And today, now that we’re a software company again, that’s more important than ever.
Just Add Compute – aka, If All You Have is a Hammer, Everything Looks Like a Nail
As consumption-based pricing becomes the norm, knowing and controlling what you’re consuming matters a lot. Many managed service providers make it “easy” by bundling hardware and software into a single offering, especially in the public clouds. But for whom is that actually easy? Well, to be honest, bundling hardware and software is much easier for the SaaS company’s developers, because when faced with customer performance expectations for complex queries, different workloads, and varying concurrency demands, the easiest way to solve performance problems is to apply more compute.
It’s also easy (and attractive) for the SaaS provider’s finance team because every hour of high-end compute means more profit. This is especially critical when the underlying infrastructure in the black box bundle is a high percentage of Cost of Goods Sold (COGS) because it is not owned by the managed service provider.
But You Lose Pricing Transparency and Control
Elasticity – especially the concept of auto scaling – is a very real value for customers using SaaS offerings, but not if it’s used instead of core optimization within the software itself. If auto scaling doesn’t allow you to specify the instance types and numbers that you’re willing (and able) to pay for, then you might indeed get the best possible performance due to the highest possible compute. But you’ll also get the highest possible invoice at the end of the month.
To be clear, ensuring the highest possible performance for predictive and proactive analytics is critical to meeting business needs, but most companies also need to accurately forecast their monthly and quarterly spend. The key to successfully balancing these conflicting demands is pricing transparency and control.
Take Back Your Performance, Flexibility, and Control with Vertica Accelerator
Vertica Accelerator is continuing proof of our commitment to delivering a unified analytics platform with the performance, flexibility, and control that our customers expect. Vertica Accelerator is a software-only SaaS offering that delivers all the query performance, concurrency, and workload isolation that our customers need while allowing them to have full and direct control of their AWS infrastructure costs: They can leverage their reserved instance types and committed contract spend. Vertica Accelerator offers both auto scaling and auto scheduling, but both provide the transparency and control of which instance types will be added to the cluster and how many.
Vertica developers love the challenge of crafting and testing every line of code to ensure that Vertica delivers the most optimized, efficient, and performant analytics platform on the market. They believe in the ongoing mission of a software-only strategy that has built powerful trust with our customers.
Would our finance department like to make money on underlying hardware? Maybe, but we’re not giving them that choice.