facebooklinkedinrsstwitterBlogAsset 1PRDatasheetDatasheetAsset 1DownloadForumGuideLinkWebinarPRPresentationRoad MapVideofacebooklinkedinrsstwitterBlogAsset 1PRDatasheetDatasheetAsset 1DownloadForumGuideLinkWebinarPRPresentationRoad MapVideo
Actian Blog / Google GKE Containers: Breakthrough Technology for Cloud Data Warehouses

Google GKE Containers: Breakthrough Technology for Cloud Data Warehouses

GoogleGKEBlog1 Image

Actian Avalanche on Google Kubernetes Engine

Google Kubernetes Engine (GKE) is now running containerized versions of Actian Avalanche cloud data warehouse, a high-performance hybrid cloud data warehouse designed to power an enterprise’s most demanding operational analytics workloads. Why is that important? Because now an enterprise can deploy one of the world’s most advanced data warehouse systems in about five minutes—a fraction of the time it takes to deploy in other cloud environments.

Let’s break this down a bit. The Google Kubernetes Engine—GKE—works with containers, which are, effectively, self-contained, pre-built component images. That’s important because in other, non-containerized, cloud environments data warehouses are typically deployed by running a series of scripts and/or REST API calls that build out each component from a base OS VM. In those scenarios, every component needs to be installed and configured—in sequence—so building out a complete cluster can easily take 25 minutes or more. That’s not a huge amount of time if you are expecting to set it and forget it, but in the age of DevOps there’s less and less setting and forgetting. The needs of a DevOps team change constantly, and in such a dynamic environment the need to reconfigure and redeploy—at 25 minutes per pop—can quickly become a source of real frustration. It’s also worth noting that a 25-minute projected deployment time assumes that everything runs without incident, and that may not always occur. The sheer number of operations that need to be executed in order to build these highly complex systems increases the possibility that something will not go as planned at some point in the process. There are lots of dots to connect, and each connection presents a point of vulnerability where something could go awry. The more you need to iterate, replicate, and expand deployments over time, the greater the likelihood that something is not going to go the way it should go and you’ll spend far more than 25 minutes trying to work out why.

Containers, in contrast, obviate the need to run these complicated setup procedures—because they have already been run and the dots connected when the containers were built. That’s right: it’s as though someone else ran through all the scripts and captured images of what a fully deployed Avalanche instance should look like—and then froze these images in a form that could be used and reused anywhere. Those pre-built images are the containers, and once built can be deployed quickly on Google Cloud via GKE.

In fact, it’s not even as complicated as deploying the containers via GKE. All an organization needs to do is select Google Cloud as the target when deploying an Actian Avalanche cloud data warehouse. Avalanche invokes GKE to do the work of deploying the containers for you and within minutes you’re up and running with a world-class data warehouse.

Making the Most of the Google Cloud Infrastructure

That brings us to the second part of why it’s great to run Actian Avalanche using GKE. Actian Avalanche is designed to make optimal use of the compute resources at hand. The more CPU power and RAM one can configure in an Avalanche cluster, the more performance you’re going to experience. While that may be true for many systems, when it comes to the cloud, distinctly different infrastructures can be implemented. And while the question of which cloud vendor has the most performant infrastructure will vary from one investment cycle to the next, users of Google Cloud can take advantage of more readily available offerings with advanced, high-performance CPU/memory configurations than found on alternative platforms, and that can be crucial in certain business scenarios where speed-to-insight is critical. The whole physical infrastructure—not just the CPUs, but also the storage and network infrastructure on which GKE itself runs—enables Avalanche to take advantage of CPUs with larger on-chip cache and faster RAM, which it has been designed to leverage. This more innovative cloud infrastructure makes it easier to access more of the processing power than in other cloud offerings.

The containerized architecture that GKE is managing is important here: containers are largely agnostic when it comes to the underlying machine hardware, which means that a containerized deployment of Actian Avalanche can easily take advantage of new hardware as it becomes available in Google Cloud. Conversely, an environment where Avalanche—or Snowflake or any other cloud data warehouse—is constructed without the benefits of containerization, will be more tightly tied to the architecture of the VM upon which the cluster components are running. Because an organization can easily subscribe to Google Cloud services that are configured to extract the highest performance achievable from the most current CPU and memory technologies, Google Cloud and GKE make it significantly easier to build a solution that will enable Avalanche to operate at peak performance.

Given the more optimal infrastructure provided by GKE in Google Cloud, it’s not surprising that provisional benchmarks conducted by Actian show Actian Avalanche on Google Cloud delivering a 20% throughput improvement on average when compared to alternative cloud platforms. For those organizations looking for the data warehouse that delivers highest performance and throughput from the cloud, Actian Avalanche on GKE presents a clear winning choice.

More advantages arising from running Avalanche on Google Cloud

Does Avalanche gain other advantages from running on GKE? Yes, but we’ll flesh those out in part 2 of this blog. For teasers, though, let me say this: Anthos and security. We’ll say more about each in future discussions about Google Cloud and Avalanche. For now though, suffice it to say that there is an early adaptor program for Avalanche on Google that will enable you to kick Avalanche’s tires yourself and see how it can meet your pressing operational analytical workload needs more effectively than ever.

Give it a shot and see if you are moved by the power of Avalanche on GKE.

About Jeremy Hankinson

20 Year Actian veteran. Senior software engineer for Actian Vector with fingers in more pies than fingers. Technology enthusiast. Lapsed physicist. Embattled Father. Avid snow boarder. Englishman living in California.