Product

Rubrik Datos IO Streaming Recovery: Near-Zero RTO for NoSQL Databases

Rubrik - Rubrik Datos IO Streaming Recovery: Near-Zero RTO for NoSQL Databases - Rubrik Datos IO Streaming Recovery: Near-Zero RTO for NoSQL Databases

Backup and recovery has been overlooked as backwaters of enterprise data. Innovations were incremental at best until the arrival of Rubrik, the company that led the world to rethink backup and recovery as the frontier of data management for cloud-connected enterprises.

Data protection has long been a pain point for enterprises across industries. Applications built on top of traditional databases like Oracle, Microsoft SQL Server, or IBM DB2 are generally stateful end-to-end. This means that whether you are recovering from media failures, site loss or user errors, you must first shutdown the application, recover the backend database, and reconnect your application before consumers can access it again.

However, things are different in the world of next-generation applications that are deployed on modern distributed architectures such as NoSQL databases (MongoDB, Apache Cassandra, DataStax, CouchBase, Apache HBase, and many others). Shared-nothing architectures, eventually consistency nature of write operations, and scale-out capabilities force us to rethink NoSQL backup and recovery. But probably the most important aspect is the shear scale of data sets that are stored and maintained in these modern NoSQL databases.

These databases power a number of modern enterprise use cases, including Internet of Things (IoT), security/customer real-time analytics, customer 360 experience, and cloud-scale content and document applications. As a result, shutting down the dependent applications for NoSQL database recovery is like shutting down the business temporarily.

Additionally, applications powered by distributed NoSQL databases are stateless in most cases, meaning they are 100% dependent on backend NoSQL data sources for all of the data and metadata needed to keep the application alive! How do we solve the application-consistent backup and durable recovery problem for such business-critical customer-facing applications?

Enter Streaming Recovery from Rubrik Datos IO, a technology introduced in our 3.0 release that powers recovery use cases such as orchestrated restores, data mobility, and test/dev refreshes.

Before we talk about recovery, it’s important to note that all NoSQL database backups performed by Rubrik Datos IO are application-consistent, bringing consistency to the inconsistent world! Application-consistency needs to be established by backup solution in this new world of non-relational NoSQL databases because upon recovery, we want to ensure that all database nodes see a coherent view of the data, avoiding time-consuming database repairs. The longer it takes to repair, the worse the RTO. So, repair-free recovery helps us drive extremely fast RTOs.

Unlike traditional data dump methods that require all data to be reloaded before applications can be made meaningfully online, streaming recovery lets your applications use the NoSQL database while a restore is in progress. This means collections in MongoDB or column families in Cassandra that are responsible for key application data are online in minutes, while those serving fulfillment and supply chain may take back seats. Furthermore, all apps are available even when data is just starting to load into the target database.

In addition to increasing business uptime, Rubrik Datos IO’s approach to backup and recovery enables users to retain full control via granular policies.  Whether you are recovering under a terabyte due to a bad piece of code or recovering an entire database after a major site-wide malware attack, your applications can become online as soon as your business-critical data is streamed back onto the data source.

While superior RTO is a key business value of Rubrik Datos IO streaming recovery, it provides additional ease-of-use benefits for customers. For example, users no longer need to stage large data sets on production systems for reload or reserve capacity in nodes to accommodate restores. It’s also no longer necessary to have the same topology at the target database.  These are just some of the ways that streaming recovery reduces costs while increasing operational agility.

With Rubrik Datos IO live streaming recovery, we’re bringing near-zero RTOs to NoSQL databases! To learn more about our newest innovations, register for our webinar: Expand Rubrik Data Protection & Management to NoSQL Databases.