Unlock Data Analytics for Dynamic Workloads with Vertica 9.1

This blog post was authored by Sanjay Baronia.

Today, cloud infrastructure has made it easier for organizations to consume services and deploy business applications with a pay-as-you-go, OPEX model. This provides a number of incentives to move data to the cloud, especially for variable workloads and use cases that require heavy compute for finite periods of time. This is particularly attractive for startup businesses and smaller teams that need to derive insights from data but don’t yet have capital outlays to build out the data centers required to handle peak compute and storage needs. If this scenario sounds all too familiar, then the question your organization should ask itself is, “How do I take advantage of cloud economics without being penalized for fluctuating requirements and dynamic workloads?”. The answer is the separation of compute and storage.

Our latest 9.1 release features Vertica in Eon Mode, enabling organizations to optimize infrastructure costs and simplify operations for their Vertica cloud deployments on Amazon Web Services (AWS) by separating compute resources from data storage. This architecture enables you to rapidly scale your compute cluster to meet dynamic workloads, independent from the storage you need to house growing data volumes. When analytic activity is low, simply scale down your cluster to reduce costs. When activity picks back up, you can scale up your cluster to keep pace. Cloud infrastructure costs are now tied directly to business value.

And, if you were to have a fixed workload or a need to deploy Vertica in a hybrid or multi-cloud environment, you can always use Vertica in Enterprise mode. This is the traditional Vertica mode, which utilizes a fixed set of compute nodes with direct attached storage, delivering blazingly fast and an unchanging pace of analytics for your business.

Vertica in Eon Mode Architecture

How does this new architecture work? Vertica in Eon Mode (see figure) leverages a single, durable communal storage location, such as Amazon S3, with additional cache copies of the data in ephemeral storage called a “depot”, for performance. Together the communal and depot storage provide the durability and blazingly fast performance you expect from Vertica, while concurrently enabling rapid elasticity to reduce your infrastructure spending. Adding a node or recovering a node is quick and easy, as there is no cluster rebalancing required. Vertica in Eon Mode also enables linear throughput scaling to meet query performance requirements with the simple addition of more nodes to the cluster.

Use Cases for Vertica in Eon Mode

Vertica in Eon Mode is applicable to a number of analytics scenarios, where workloads have to scale up and down to meet certain business objectives. The following is just a subset of the many analytics use cases that can benefit from the separation of compute and storage:

1.Experimentation and Discovery for Optimized Device Setting – HVAC manufacturers need to run periodic maintenance tests by pushing an update to edge devices, collecting a large amount of sensor data, and quickly identifying micro patterns before pushing out optimized settings. With Vertica in Eon Mode, HVAC manufacturers increase the compute and storage resources seamlessly to accommodate for increased activity.

2.Engagement analysis for new product introduction – Gaming companies must increase analytical capacity during major launches or tournaments to evaluate the success of a new game or provide customer insights during a big tournament or special event. With Eon Mode, gaming companies derive just-in-time analytics on critical feedback by event – in time to make modifications that appeal to their community — as opposed to the ultimate launch, when it may be too late.

3.Analytics for seasonal sales patterns – Retailers have variable analytic workloads due to high-volume sales seasons or to close the books at the end of the month. They need maximum dashboard performance regardless of the number of concurrent users, so that key members gain access to important metrics. Eon Mode ensures maximum dashboard performance at peak times regardless of the number of concurrent users, so that key stakeholders gain access to important metrics irrespective of the load on the analytics platform and the seasonality.

4.Project-Based Data Science – Contract data scientists are increasingly used for project-based work, where they ramp analytics projects up and ramp down to meet specific client deliverables and schedules. Eon Mode empowers contractors with more flexibility to address client needs faster without the fear of costly overprovisioning.

5.Data Science and Engineering Team Collaboration – Machine Learning models are often built in the cloud before ultimately being deployed onto a much larger, production, on-premises analytics platform. Eon Mode provides a “hybrid” bridge that allows data science and data engineering teams to collaborate on a best-of-both world’s approach to machine learning model development, testing, and deployment.

How Can I evaluate Vertica in Eon Mode?

Getting an Eon Mode database up and running is fast and easy. Don’t take our word for it. Try it yourself!

All current Vertica license holders are able to use Eon Mode if they upgrade to 9.1. Since Eon Mode is a Vertica feature (not a separate product), anyone with a valid Vertica license can deploy their database in Eon Mode. It’s that simple. And, if you do not have a license then you can spin up the Vertica Community Edition and evaluate Eon Mode for free, right from the AWS Marketplace.

For information about creating a Vertica cluster on Amazon Web Services (AWS) and creating a cluster using an AMI, see Vertica on Amazon Web Services in the Vertica product documentation. For more details about creating an Eon Mode database, see Using Eon Mode in the Vertica product documentation.

When you are ready to purchase, you can either contact us or if you prefer usage-based consumption, then Vertica is also available as an hourly paid subscription from the AWS Marketplace.