Databases technologies and products in the NoSQL database category have now been around for a decade, or at least the better part of one. And in the ensuing years, there's been a combination of consolidation and maturation. There's also been a de-emphasis of the products' more "militant" design philosophies -- which is why most vendors in the space no longer even use the "NoSQL" moniker .
Stacked in Cassandra's favor One of the stalwarts of the "database genre formerly known as NoSQL" is Apache Cassandra. Its community's major commercial entity, DataStax, is a stalwart as well, and DataStax Enterprise (DSE), the company's premium/commercialized distribution of Cassandra, has itself been around since 2011. Today, the company is releasing version 6 of this flagship product.
What's new? DSE 6, rather than adding only raw new capabilities, is instead adding smarts that make the existing capabilities faster, easier to deploy and easier to manage. In short, it's an Enterprise fit-and-finish release that provides evidence that the NoSQL genre is taking seriously the notion that products need to be easy to use and manage, or their adoption becomes cost-prohibitive at large organizations.
On the administration and management side, DSE 6 is adding features to make repair operations a cinch; ease resource contention on individual nodes in a DSE cluster; and make rollout of patch upgrades a friction-free operation. And under the DSE Advanced Performance banner, DataStax is adding better performance, a new data bulk loader and important enhancements to both DSE Analytics and DataStax Studio.
Node care Cassandra offers a "masterless" architecture, which means no single node is indispensable to the cluster. This enhances fault tolerance, availability and resiliency, but it has also created a burden: node repair operations, necessary to support the architecture, have been manual and arduous. In DSE 6, a new "NodeSync" feature makes node repair a completely automatic and implicit operation.
Effectively, with the version 6 release, DSE becomes self-repairing. And while OpsCenter will still provide an interface to monitor the operation, such monitoring is purely optional, as the operation is now autonomous. Features like this one make it crystal-clear that DataStax understands Enterprise pain points and wants to address them head-on.
Another proactive and automated feature in the care and feeding of DSE nodes is "TrafficControl." This feature will address Java Virtual Machine stress and other overload effects when too many concurrent requests are routed to a particular node in the cluster. The feature orchestrates the requests by queuing them, thus allowing the node to process them in an orderly fashion.
Upgrades happen In the age of continuous improvement and continuous deployment, product updates ship continuously too, and applying patch updates to DSE has been a labor-intensive operation. But a new Upgrade Service, built into OpsCenter, will allow DevOps personnel to manage this process much more easily.
Upgrade Service will automatically notify customers of upgrade availability and will download the upgrade bits automatically as well. It will then allow patches to be applied to nodes in the cluster, one at a time, allowing the cluster to stay up and running. Essentially, Upgrade Service makes patch updates an non-disruptive, online operation. And, in the future, Upgrade Service will handle deployment of bigger releases as well as patches.
Capabilities matter too As nice as it is to for a product's management to be streamlined, it's also important to improve the core product itself, and DataStax has in fact done so with DSE 6. For example, DSE Advanced Performance will allow the DSE product to provide an average of twice the performance of vanilla open source Cassandra, according to DataStax.
The new Bulk Loader will allow highly accelerated ingestion of data, from external delimited or JSON text files, into a DSE database. Bulk loader is comparable to utilities, like SQL Server's BCP (bulk copy program) and Oracle's SQL*Loader, that have been available on relational platforms for multiple decades. But let that not be a disparagement. For a newer product to backfill essential functionality that older products have long supported is a sign of sensitivity and maturity. Enterprise customers rely on these tools and even newer competitors need to provide parity.
Moving on, DSE 6's new Advanced Apache Spark communication layer will optimize performance of DSE analytics, providing up to 3x query performance, according to DataStax. And DataStax Studio will now provide a notebook interface for execution of DSE Analytics queries -- rounding out the support for CQL and graph queries that DataStax Studio already provided.
Watershed? There's more to DSE 6 than I've covered here, including a new AlwaysOn SQL feature, enhancements to both DSE Advanced Security and DSE Graph, as well as support for Spark Structured Streaming. DSE 6 is clearly a big release.
DataStax's customer count is now north of 400. And while that may seem small compared to the incumbent players, in my book, it's a number that signifies critical mass. 400 customers is a rite of passage, and companies that go through it tend to be mature enough to prioritize addressing customer pain points and go after increased adoption in a sober, methodical fashion. The nature of enhancements delivered by DSE 6 is consistent with that observation.
While some of the newer database companies have faded, struggled or bought time with more funding, it's clear that DataStax is focused on delivering an Enterprise product that is not only innovative, but performs well, can be operated in a low-friction manner, and generally caters to the needs of big companies who want database platforms that scale in a repeatable and manageable fashion. That's what makes a stable industry.