Since the dawn of the information age, replication for disaster recovery has been business as usual for the data center.
In more recent times, many business intelligence tools have emerged that also use replication. But, these processes are relegated to the night shift, meaning done outside of the normal business day.
Fast forward to today and this approach doesn’t stand up to the 24×7 mentality of the modern world and the global, virtual data center.
Add to this never-sleep mentality, rapid data growth and new uses for (big) data warehouses and you can see the need for replication technologies that push the boundaries of data recovery and also feed the analytics engines that drive business intelligence decisions.
Getting there means embracing the not always obvious reasons for replication and retooling and re-organizing to deliver self-service models and to scale.
Key to a successful journey is taking an application and replica management approach to replication.
Thinking Outside of the Box
Effective strategies for business continuity are built around replication tightly integrated with federated application infrastructures. The rapid resumption of key business processes from any type localized or regional disaster is the most often cited use case. But, disaster recovery is not the only reason you replicate data.
Let’s step through these three (3) not so obvious reasons for replication.
Recently, we addressed the emergence of email as a business document relevant to litigations and subject to regulations. Replication and granular recovery of Exchange email data and potentially other electronic data helps you and others meet the archived communication requirements now mandated in many jurisdictions and industries.
EMC Replication Manager creates consistent copies of critical applications like Exchange for recovering important documents, including for compliance. Other best practices include the ability to provide granular restoration, including single virtual machines and single Exchange mailboxes with the Replication Manager extension, EMC ItemPoint, to meet compliance needs, as discussed in the recent post.
2. Big Data Analytics
Another reason to replicate can be found in physical and virtual infrastructures that leverage new, emerging technologies to drive business intelligence. Replicating data stores can be difficult when moving large amounts of data over limited-bandwidth networks. Fortunately, Big Data appliances such as Greenplum can load data at very high speeds, avoiding the need for overnight or weekly snapshots.
Restart can be a challenge in these deployments, but integrations with enterprise-class replication solutions like EMC SRDF can recover large amounts of data in almost no time. Add EMC PowerPath for data path optimization and fast performance and recovery times are quicker with less impact.
While admittedly more obvious than the first two reasons, how you manage copies of data for purposes other than backup and recovery is evolving from a data focus to the application level. These event-driven systems have extreme flexibility and can initiate business processes based on certain data conditions, creating unparalleled business agility.
For example, at the most basic level, a mobile customer service representative accessing a customer account record may see a customer’s eligibility for an upgrade when taking a payment over the telephone. The call center employee can initiate a sale, with this information.
Because the data is a current refresh from a production replica, the timely retrieval of the most current information happens instantaneously, providing the carrier with an instant competitive advantage over possibly another carrier offering a similar service.
Another example and closer to home is EMC’s own journey to better repurposing. EMC’s data warehouse had grown to unwieldy proportions. EMC IT used replication to address the business intelligence challenges.
What about disaster recovery?
Business as Usual
A Gartner Group report once cited less than 1% of data outages as caused by actual disasters—natural or otherwise and 15% caused by infrastructure failures. Most outages (85%) were planned or scheduled within maintenance windows. This could be data mart refresh, an environment upgrade, or data migration.
That said, with the growth of virtualization, cloud, and Big Data, there is extreme sensitivity to data loss or downtime in these massive virtual infrastructures. It’s critical that your recovery scenarios encompass any potential outage scenarios and leave no vulnerability exposed.
But, restart of a failed application is not a slam dunk, especially in federated environments which can cause big headaches in your recovery scenario, and especially if your data is only crash consistent.
It’s important to pay attention to what is happening up the stack, ensuring replica copies maintain strict application consistency and data integrity throughout the replication and recovery process. To this end, technologies such as EMC RecoverPoint EMC RecoverPoint working with Replication Manager integrate local and remote replication with your core applications in a VMware virtual environment.
Top disaster recovery strategies are those that leverage replication integrated with business-critical applications. Application-aware replication also plays well into the repurposing use case described above.
What’s on the Horizon?
Administrators are using solutions such as RecoverPoint and Replication Manager, as well as snaps and clones for data repurposing and recovery. But, when talking to customers, I often hear that they are looking to make replication an integral part of their overall data center strategy.
To be effective, it is very important that you allow application owners to manage replica copies of data in a self-service environment while administrators define service levels. This empowers application owners to manage their own replicas on demand and select services based on pre-defined policies.
To get there, enable the IT administrator with tools that provide centralized control for defining service tiers, mapping the service tiers to the underlying replication technologies (whether point-in-time or mirror-based replication or snaps), and then allowing the application owners to assign service levels to applications.
There are technologies today that equip you to take full advantage of replication as part of your overall service offering and get you on the road to these less than obvious reasons, but advanced replication use cases. These technologies might not address all Big Data needs, but most assuredly they’ll get you to compliance and application-level repurposing.
Keep tuned to this blog for more news in this direction.