NuoDB 4.0 improves cloud-native database capabilities

NuoDB targets multi-cloud

NuoDB is continuing to aggressively pursue the multi-cloud strategy the vendor started discussing with its version 3.0 release, said Merv Adrian, research vice president for data and analytics at Gartner.

Gartner surveys show that 80% of firms in the public cloud are using more than one, Adrian noted.

In that light, NuoDB's strategy "may be of interest to larger firms interested in being able to operate consistently across such fabrics," he said. "It will simplify the consistent application of policy for compliance requirements."

NuoDB's target market is customers that are running applications on Oracle or SQL Server on premises and are committed to either moving to the cloud or to a digital transformation to update to a modern architecture, said Ariff Kassam, CTO of NuoDB.

"As part of that migration or that update of architectures, the architects are looking at not only the application tier, but as well as the data tier, and they're looking for a database to sustain them for the next five or more years," Kassam said.

The NuoDB architecture can enable the database to look like a SQL server or Oracle database from an application perspective, but inside the system has been optimized to take advantage of modern deployment targets, such as Kubernetes and multiple public cloud providers, Kassam added.

Hybrid and multi-cloud computing

In the NuoDB 4.0 update, Kassam said that being cloud-agnostic and cloud-native were two key areas of focus. Cloud-agnostic is also often referred to as multi-cloud and is an approach in which a technology supports operation across multiple public cloud providers. The cloud-native term is often used in reference to Kubernetes, which runs in the public cloud and on-premises and is often used to help enable hybrid cloud models.

Kassam said that previously NuoDB supported AWS as a public cloud provider. Now, Nuo 4.0 is certified for Google Cloud Platform and Azure as well.

Adrian said he is optimistic about the market opportunity for Kubernetes and cloud-native optimized databases.

Most database management system vendors are talking about supporting Kubernetes because they all see it as a critical checkoff item competitively, he said.

Databases that build horizontal resiliency into their architecture are better suited to leverage this new wave of interest in container technology, and NuoDB can get an edge here early if this delivers as promised.

In addition, Adrian noted that if lifecycle management and good operators are in play, it means easier deployment with a governed self-service approach. That means organizations can be more agile in deployment and experience less reliance on calling data center specialists to add or change deployments.

"Databases that build horizontal resiliency into their architecture are better suited to leverage this new wave of interest in container technology, and NuoDB can get an edge here early if this delivers as promised," Adrian said. "How much the market wants it today is a little hard to gauge."

In terms of cloud-native, Kassam explained that before NuoDB 4.0, the vendor supported containers and Kubernetes, specifically via Red Hat's OpenShift Kubernetes product. With the new NuoDB release, the support is for Kubernetes Operators, which is an approach for packaging and deploying an application into a Kubernetes cluster. The move to further enable NuoDB for Kubernetes is also about a shift in perception and usage for Kubernetes.

"A lot of people have been using Kubernetes to deploy stateless applications, where the application is in containers, but the database has been outside, in virtual machines or physical servers," Kassam said. "While that works, it's not ideal, as there are security and operational issues in dealing with two sets of management environments."

Kassam said as recently as a year ago, some were pessimistic about whether stateful applications could run in containers. He said much of that pessimism has evaporated the Kubernetes open source project and storage vendors have done a lot of work in stateful applications.

"I think now people are accepting the fact that you can run stateful applications within Kubernetes," Kassam said. "There's more work and the architecture is a little more complicated, but the pieces are coming together."

How NuoDB 4.0 works with Kubernetes

Kubernetes is a highly modular platform with multiple ways to accomplish any given task, such as connecting a NuoDB 4.0 database.

Kassam said NuoDB aims to give users choices on how they want to deploy a database. One option is to use a Kubernetes approach known as local persistent volumes, which provide an attached data volume to a set of containers running in a node.

Another approach is to plug into the Container Storage Interface through third-party data storage providers such as OpenEBS, Linbit, StorageOS and Portworx.

NuoDB admin

Among the major new features in NuoDB 4.0 is the NuoDB Admin management interface, which integrates a REST API for improved extensibility.

"One of the things we struggled with a couple years ago, as we were moving to containers, is we needed different API access for how to deploy things either on premises, the cloud or in a container environment," Kassam said.

Indexing improvements

At the core of most databases is some form of index, which provides the basic data structure that enables data location and access. In NuoDB 4.0, online index creation and expression-based index capabilities have been added to the database.

The online index creation capability means that NuoDB 4.0 creates an index without having to lock the table or take the table offline. He added that Oracle has had that capability for some time and it is a capability that NuoDB customers were requesting.

With the expression-based index capability in NuoDB, the database gets a performance boost that organizations will notice, Kassam said.

He said that without an expression-based index, the database index includes the values of one or more columns and any queries that use those columns can use the index. For a query workload that uses functions for accessing data -- for example, taking a date from a timestamp value or calculating some sort of function off of multiple values -- those generally did not have the ability to use an index, so they were slower in previous versions of NuoDB.

"Now, with expression-based indexes, customers can speed up those queries that use general functions for querying," Kassam said.

He added NuoDB plans to continue to improve its support for cloud-native deployments as Kubernetes continues to evolve.

Also, he said, "we continue to look at improving performance, improving our optimizer and we're looking at a new vectorized execution engine and some machine learning capabilities in the next iteration."