At our recent webinar, “NoSQL vs. NewSQL: Evaluating Database Technologies for 2019”, we received a number of questions from attendees about VoltDB’s capabilities and architecture. From orchestrating VoltDB with Kubernetes, to comparing VoltDB to TimeScaleDB, here’s a recap on the key questions asked during the webinar.
For those who missed the webinar, you may view it on demand.
Editor’s Note: Unless otherwise noted, the question is coming from a webinar attendee via our chat functionality during the session. Answers were provided by VoltDB’s Chai Bhat and Madhup Mishra. These answers have been edited for clarity and grammar.
Q: Can VoltDB run analytical queries?
Yes. VoltDB has what they call HTAP. It’s a hybrid transactional, analytical processing operational database so you can actually run analytical queries. In addition to your traditional OLTP, you can do all that queries as well. Good question though.
Q: Is VoltDB available on AWS?
Yes. VoltDB is available on Amazon Web Services. We are in the AWS Marketplace so you can download it there use us. Also, if you’re on the Google Cloud or on Azure, we have a container Docker image that you can use to deploy in those clouds as well, so we’re cloud agnostic. You could pretty much deploy us on any public cloud or even a private cloud for that matter.
Q: Is VoltDB certified on OpenShift?
Thank you, that’s a good question. So Red Hat has the OpenShift platform. We know of customers that do use us in that capacity. However, we haven’t officially passed the certification for that yet. We also have a Kubernetes-based implementation that we launched that is available for you guys to use, to deploy us at scale. It allows you to spin up a VoltDB cloud using Kubernetes and manage it over time.
Q: Can I orchestrate VoltDB using Kubernetes?
Yes, you absolutely can. Traditionally, VoltDB being a stateful database and Kubernetes being designed on web apps – which are dynamic – it was a tough problem to solve. But yes, now VoltDB can be orchestrated with Kubernetes.
Q: How does this scale horizontally and across datacenters?
Yes. So VoltDB does partitioning. Basically, what it lets you do, as I showed in that benchmark, is that you can actually queue; you can use multiple CPUs within a server, or across servers, to take advantage of our partitioning. So basically, commands are queued by partition and they are executed by each CPU across partitions, so each command is executed individually in each participation. And each partition maintains strong active semantics and we can replicate, not just across servers, but also across different clusters and different geographies as well. So, we have active-active disaster recovery that goes across different regions even; so, if you have a data center in Europe, you will have replicas in Europe along with your data center in the US. If the US data center goes down, you can keep running in production because you have a replica on a data center in Europe.
Q: Are there any administrative tools to manage VoltDB?
Yes. There are lots of tools. We encourage you to Google us. Go to VoltDB University and VoltDB. All the documentation is online so feel free to check us out. We do provide a VoltDB management console for those of you who are not CLI (Command Line Interface) people to take a look at the [pace] and speed and how the clusters are doing and what the configuration is. We can also be run through our SQL command operator so you can manage us via CLI as well.
Q: How straightforward is the data migration from other databases?
Excellent question, right? I have yet to meet a database company that has a smooth answer to that question. So ultimately, it comes down to how well you understand your application and what it’s written in. There are certainly tools available for you to move things over to VoltDB considering we’re SQL-based. However, because we are NewSQL and we have a partition core-based architecture, there are nuances in how your application will get charted across lots of different servers. So, as long as you understand what you’re doing, it’s something that you can get done fairly easily. We also have services and support that we offer helping customers who need some help and support along the way.
Q: How do you compare VoltDB to TimescaleDB?
So TimescaleDB is one of many time-series database options that are fairly popular these days. Time series is a subcategory of NoSQL which is what Chai Bhat discussed earlier on. You can manage time series within VoltDB because VoltDB does support time series capability. However, we’re not simply a time-series-based database like TimescaleDB.