All of us need a sense of purpose, a reason to apply our energy and passion to something that matters. I wake up each morning with a mental list of key priorities, and a longer list of other tasks that are necessary but not nearly as important. I work hard to stay focused on what really matters, but I can’t deny that I often get distracted, leading to a lot of open Skype chat windows, email drafts, and an incomplete or less than impressive outcome for the things that matter most. Except, of course, when it comes to my dog -- my sense of canine purpose is never compromised.
The Promise of Cloud Economics
Multi-tasking between low-level tasks and critical business projects is not just a human behavior – computers are forced to do the same. Traditional data center architectures that were built on servers with specific CPU and tightly coupled storage find themselves responsible for real-time, executive-level dashboard reporting, individual business unit projects, and random acts of ad-hoc querying. Data center architects are accountable to meet all business use cases and performance expectations, so, naturally, they provision for the biggest and most critical workloads. But with IT budgets shrinking and the tantalizing lure of “cloud economics” on every billboard and at every major airport, the financial flexibility to overprovision compute to protect the most important workloads is no longer possible.
Let’s give credit where it’s due -- one of the biggest and most disruptive contributions from public cloud platforms like AWS, Azure, and GCP is the next-generation data center architecture that separates compute from storage. The advantage introduced by the public cloud providers is how they provide specific compute capacity for each task while providing distinct (and low-cost) storage locations for whatever amount of data is needed. Hence the phrase “cloud economics” – buy the compute you need, the storage infrastructure you need, and only pay for what you need, when you need it.
Computation with a Purpose
But, the big news here is that this architecture is no longer limited to public clouds. Cloud economics is coming down to earth, as companies who choose to manage their own data centers or IT departments who operate private clouds also evolve their data center architectures. The ability to link specific workloads to the compute that they need when they need it is now the next-generation data center architecture, both on premise, in private clouds and on the public clouds. Computation with a purpose has arrived.
When we envisioned the next generation of Vertica, we saw the impact of computational purpose (that’s a cool phrase, right?). But we knew that architecting a solution for a single cloud platform would compromise the very foundation of Vertica – freedom from underlying infrastructure. Instead, we chose a more complicated path, but the outcome was worth it!
Right now, you’re certain I’m talking about Vertica in Eon Mode. And you’re right, but this vision began even before Vertica in Eon Mode. Vertica began this journey when we built our integration with another ecosystem you might have heard about -- Hadoop! When we built our ORC and Parquet readers, we created a Storage API that enables Vertica to interact with HDFS, which is Hadoop’s file store. When we developed Eon mode, we built an S3 implementation of the API, enabling Vertica to natively access S3 for both internal ROS files and external data lakes of ORC/Parquet. There’s a company out there in the Big Data snowstorm that can’t do that.
Vertica’s architecture is very special. It is the foundation and the differentiation that allows us to be the industry’s only infrastructure agnostic unified analytics platform for all your data, regardless of the location and format and purpose of that data, with no compromise on performance.
The Benefits of a Modular Architecture for Analyzing All Your Data
Vertica’s modular architecture allows us to extend Vertica’s Eon Mode deployment to multiple clouds and to on premise data centers, bringing cloud economics down to earth. Vertica’s modular architecture separating compute from storage allows us to extend the full suite of ANSI SQL and advanced analytical functions and in- database machine learning not only to ROS data in S3 in Vertica in Eon Mode, but also to the extensive data lakes that are growing in every organization today.
- Know and respect the role that open source formats play in an organization’s data strategy.
- Understand that multiple applications often need access to the same data.
- Realize that the “hottest” data, whether that is defined by timestamp or other specific criteria, needs millisecond attention and results.
- Compete in a market where many vendors offer analytics databases and all you have to do is put all your data in one place.
- Encounter many vendors that offer query engines, and you can use these tools on data stored in open source formats across different (but not all) data lakes.
But we don’t compete with any vendors who can do both, because there aren’t any. Vertica is both a high-performance, native columnar database AND an open format query engine. Vertica applies purposeful compute based on business needs for specific workloads and data sets with independence from the location and format of that data. All of this is possible only because the underlying architecture of Vertica was built on not only the power and performance of a columnar database but because we saw the power of computational purpose. Vertica is bringing cloud economics down to earth but with a key difference - for all the data.
Want to learn more about Vertica?
Get up and running in minutes with Vertica by the Hour on AWS or download the Community Edition directly onto whatever infrastructure works best for you. Still in the research phase? Check out our great case studies to see how other organizations are using Vertica to quickly derive insight from high volumes of varying forms of data. Or just shoot us a message.