Last weekend, I finally had the chance to watch the popular, yet alarming HBO documentary – The Inventor: Out for Blood in Silicon Valley
. If you haven’t seen it, it’s a true story that centers on Elizabeth Holmes, ex-Theranos CEO, who dropped out of Stanford to build a company that promised to revolutionize blood testing using a small amount of blood from a finger prick.
Warning – slight spoiler alert ahead.
In short, Holmes and her management team overpromised and way underdelivered. In fact, that may be the understatement of the year. In 2015, Theranos had a $9B valuation and a clear path to a successful IPO. Today, the company is defunct and Holmes and her former COO are facing lengthy potential prison terms for wire fraud.
Infatuated with Steve Jobs, the charismatic and black turtleneck-adorning Holmes had a flair for the dramatic along with a vision for a blood testing device – dubbed Edison – that dazzled advisors and investors with the potential to perform up to 240 tests (from celiac to cocaine). The laws of physics – the Edison box physically could not fit the components and instrumentation required to run all of those tests – combined with unrelenting pressure from investors to prove that the technology actually worked, ultimately sank the company.
The “Black Box” Analytical Databases Behind DWaaS
Edison got me thinking about the Data Warehouse as a Service (DWaaS) space, particularly the “black box” analytical databases behind DWaaS. There’s no doubt that DWaaS is expanding our market for organizations seeking to embark on their predictive analytics journeys with a few clicks and a swipe of a credit card. But, what are the long-term costs of DWaaS and are organizations blindly following similar false promises only to be left in the lurch when their analytical initiatives change direction? After all, in the IT industry, we know that the only constant is change.
In 2012, Amazon was one of the first DWaaS entries to market, licensing the ParAccel analytical database as the core of its system. ParAccel was eventually acquired by Actian, which was acquired by HCL last year. And like that, Amazon Redshift was born. However, it was widely known that Amazon needed to shut off many parameters of the ParAccel database to make it more stable, not to mention that there’s no separation of compute and storage architecture to address the elasticity requirements of dynamic workloads.
That same year, Snowflake Computing was founded. Its bold claim? The Data Warehouse for the Cloud. What’s even bolder? The company has raised nearly a Billion (that’s with a capital B) in investment in just more than five years in existence. Valuations are north of $3 Billion. The management team was recently sacked and replaced, as pressures mount to go IPO sooner rather than later.
But, what are the actual database capabilities behind Snowflake? Well, that remains a mystery because you can’t actually see it and, apparently, the company believes that there’s no need for you to tune the database or queries. When queries are running slowly, just submit a support ticket and with a push of a few buttons – poof! – the queries run faster and all customers are happy. Right?
Well, management and finance, may have some concerns. When they get their AWS bill, they realize that there’s a whole lot of compute behind the “easy” system to crunch all those queries. And, since it’s a cloud-only database, what happens if your organization encounters regulatory or IT changes that make moving to on-premises necessary? You’re stuck.
Do Your Homework and Remember to Keep Your Future Options Open
For SaaS-first companies, particularly start-ups, DWaaS may be a good option to get started with analytics initiatives. But, make sure to do your homework to understand what your organization actually needs from DWaaS – flexibility, predictability, speed, and a fast start? Cloud-optimized databases like Vertica
address all of those needs, while preparing you for any future pivots to your analytics strategy.
Vertica – the #1 Cloud Data Warehouse with Faster Performance at Lower Costs
Vertica is one of the first analytical databases to run on AWS
followed by support for Microsoft Azure and Google Cloud Platform. With Vertica in Eon Mode
, you also have the option to run its separation of compute and storage architecture to address dynamic workloads, while tuning for performance and ensuring complete control as your deployment options invariably change in the future. You can get up and running in a matter of minutes and choose from a variety of flexible pricing models that meet your CAPEX or OPEX requirements, including subscription pricing and by-the-hour consumption (run on-premises, on all major clouds, on Hadoop, or a combination, all with a single license).
But, don’t just take my word for it. IT Central Station just ranked Vertica as the #1 cloud data warehouse
, based on product reviews, ratings, and comparisons. All reviews and ratings are from real users, validated by IT Central Station.
And, in terms of performance for value, third-party benchmarks show that Vertica in Eon Mode is up to 63% faster than Amazon Redshift and 31% cheaper from a cost-per query standpoint
. Snowflake? Independent benchmarks also prove that Vertica in Eon Mode performed queries 40x faster than Snowflake with 20 concurrent queries and delivered 89% lower cost per query
. In fact, Snowflake was unable to complete some of the 5 TB and 10 TB tests in under two hours.
So, the next time you evaluate a DWaaS, ask about the black box analytical database behind the cloud service and make sure that you don’t get locked inside of it. After all, history has taught us that meteoric valuations and promises of IPOs can force vendors to take short cuts – cuts that bleed your budget and leave your analytical initiatives at risk in the long run.
Convenience But at What Cost?
Announcing Vertica Version 9.2.1 – Take Analytics Efficiency to the Next Level
How to Access Flexible Cloud-Optimized Analytics on AWS
Vertica’s Architectural Direction – Separation of Compute and Storage