By Edgardo Lopez, Product Manager

Turning higher availability and data integrity into cost effective DR solutions that decrease business risk and improve productivity

In our most recent discussion, we talked about how non-disruptive data migration with the HP StorageWorks SAN Virtualization Services Platform (SVSP) can significantly help firms improve data mobility and asset management with the potential of significant savings in capital expenses.  In this new discussion, we want to turn to the issues of business continuity and disaster recovery.  

Many people today confuse business continuity with disaster recovery. Business Continuity ensures that enough resilience (or high availability) has been built into the infrastructure to negate single points of failure within a single site to ensure "business as usual" availability. Disaster Recovery adds recovery after loss of the primary site. Both need to ensure that the Recovery Point Objective (RPO), the point where the application data can be successfully recovered and the Recovery Time Objective (RTO), the point where the application is live again, are within the guidelines of the business.

So, what is the problem?

In the U.S., the 9/11 events have made disaster recovery a major concern for businesses of all sizes. But, while large business can afford the consultants and testing required to build a complete data recovery plan, small and medium size businesses are often challenged by the finances and the choices available. Even though the business risk involves loss of data and applications, not all business are able to deploy an adequate DR solution.

Traditional choices include clustering, server-based data mirroring solutions, and controller-based mirroring software.   However, all of them introduce shortcomings (see HP SVSP: Technology Overview and Use Cases - Note #1 below - for additional details) that negatively impact the complexity and cost of the solutions.  

Finally, the DR solutions and schemes should conveniently allow for non-disruptive testing of the recovery scheme, to make sure the business in not in jeopardy, should it actually need to use the DR solution.   Many organizations however, will give up the DR testing or perform minimal testing simply because of the disruption, complexity and cost of these tests.

It is in this context that we want to introduce the value of a DR solution with an SVSP-based storage infrastructure.   

The SVSP provides both synchronous and asynchronous data mirroring capabilities that facilitate instantaneous service resumption after storage, site or regional disasters.  Its ability to support heterogeneous storage makes ideal for enabling greater flexibility in the choice of storage hardware at the primary and remote sites, to better optimize solution costs.  It also provides space efficient copy services, including snapshots, which will keep the system on-line and fully available during backups, allowing restoring of the entire application in minutes rather than hours.  Centralized management of the recovery operations reduces administrative tasks, improves productivity and minimizes errors. 

The SVSP remote mirroring solution uses a snapshot-based technique, where only the differences between the snapshots are transmitted. This avoids the need for very expensive communication lines between the locations. It is possible to use common (or readily available) and low-cost communication lines, rather than higher-bandwidth higher-cost lines, further improving the economics for small and medium size businesses.

About Data Integrity (Note 2):

Data Integrity is another important requirement for remote replication solutions.  There are multiple aspects:

1. Consistent copies when multiple volumes are involved
2. Handling intermittent failures. Disasters often unfold over a period of time
3. Impact of data corruption on the total recovery time. 

If the wrong recovery choice is made then the problem could be compounded or a total data loss could result.   The SVSP provides a very flexible set of capabilities to handle these issues and enable firms minimize business risk.

Let's talk about the first issue.  The SVSP asynchronous mirroring supports consistency groups.  These are groupings of volumes to enable the creation of consistent Point in Time (PIT) Copies across multiple volumes. This function is important because it aids good application design without sacrificing data consistency requirements. For example, in a database environment, the database and log files can be placed on separate volumes, and the consistency group function will allow the creation of consistent PIT Copies across all databases, log files, and volumes. This is critical because a failure to capture all of these components at the same point in time would leave the database application in an inconsistent and therefore unusable state. Consistency groups are therefore essential in large application environments. They can also contribute to more efficient management of application servers and enable PIT Copy automation through scripting.

Now let's turn to issues 2 and 3.  The HP SVSP provides solutions to these two problems. By using PIT Copies, there are many points in time in which to check for consistent data sets. Once a consistent state is found, the database can be brought online and if possible rolled forward. Application integrated PIT Copies are replicated to the DR site and the consistent data can be brought online either in the DR or Primary site very quickly without the need to recover from tapes. To enhance the rapid recovery asynchronous mirror failback (HP SVSP Continuous Access) facilitates true DR testing and enables replication of just the block changes from the DR site back to the Primary site without a complete block level synchronization of all data. This saves money otherwise spent on communications link bandwidth to support data change rates.

For more information on the SVS, please visit us at:  www.hp.com/go/svsp

Note #1: "HP SVSP: Technology Overview and Use Cases" and HP white paper Paper URL: http://h20195.www2.hp.com/v2/GetPDF.aspx/4AA2-3242ENW.pdf

Note #2: "Rapid Application Recovery with the SVSP: Using PITs and DR to meet today's stringent recovery objectives", and HP white paper Paper URL: http://h20195.www2.hp.com/v2/GetPDF.aspx/4AA2-6392ENW.pdf  

Tweet this!