VPLEX – Operational vs. Business Benefits

VPLEX – Operational vs. Business Benefits

Posted on March 2, 2012 by TheStorageChap in Data Center, Federation, Virtualization, VPLEX

Whilst it has been a busy few weeks from a work and travel perspective, with another busy week ahead (I am on a plane to Dubai as I write this), I am hugely encouraged by customer demand for the functionality provided by VPLEX. Last week I was in South Africa presenting both RecoverPoint and VPLEX to several customers and Service Providers.

One of the barriers to sale that I sometimes hear and in fact did hear from one of the Service Providers, was that customers do not have a requirement for the increased availability that VPLEX can provide, “..they are happy with Active/Passive replication.”

Putting aside the fact that I do not believe this, based on having spoken to many customers, and in fact spoke to two customers during the same trip who did want to increase their datacentre availability, I drilled down a little further on their statement.

The crux of the matter is the cost to deliver such a service and what premium the consumer is willing to pay for that additional availability. A large proportion of the benefit derived from VPLEX is the reduction in planned downtime. In a Service Provider environment, in which they are hosting the customer’s primary infrastructure, the issue is should the customer have to pay for something that they probably already have an expectation that the Service Provider will deliver? There is no doubt that functionality such as non-disruptive data mobility and the ability to non-disruptively move running virtualised applications from one datacentre to another datacentre has operational benefit for the Service Provider, in terms of being able to offer that customer additional application uptime, but the problem is that it is exactly that, an operational benefit.

The other benefit of VPLEX is the ability to simplify application failover between the datacentres, but this simplification is again more likely an operational benefit for the Service Provider who is hosting the application service rather than the end-user consumer. The end-user consumer has an expectation that in the event of a datacentre disaster their application will be restarted in a timely manner at the remote site.

Service Providers need to be able to bill their customers for all of their expenses. Today if it costs x per TB of storage for the consumer, a replication service tends to be sold for

multiplied by two (source and target storage) + 30%-40% premium for replication technology.

This model tends to hold true for large enterprise IT Departments also, as they tend to operate an in-house Service Provider model that needs to be competitive.

VPLEX needs to be considered separately from both a Business and Operational Benefits point of view. Whilst the cost to implement VPLEX may be more than a standard replication solution, the potential decrease in the cost to create and administer application and storage services also needs to be factored into the finance equation.

Active/Passive replication service

Let’s take a standard virtualised server environment running VMware, using SAN storage, replicated between two storage arrays using array based replication and integrated with VMware Site Recovery Manager. For most Service Providers/IT Departments they will standardise on an array manufacturer and replication technology to allow for some sort of operational efficiency (it is also a requirement for the SRM SRA), even if the cost per TB of storage is potentially higher than from other vendors.

With this service the consumer gets active/passive failover but at what cost to the Service Provider/IT Department?

  • VMware Site Recovery Manager License.
  • Same array model in both sites.
  • Replication software license.
  • Double the capacity, but only the I/O utilization of one array.
  • Operational complexity of ensuring cooperation between SRM and storage array replication.
  • The loss of a storage array results in a requirement to failover application services.
  • Manual intervention to failover in the event of an issue.
  • Downtime for datacentre maintenance.
  • Downtime for storage refresh.

Active/Active service

In a similar service based on VPLEX the consumer gets active/passive application failover, but on an active/active storage presentation that also enables non disruptive movement of VMs between sites, enabling the Service Provider/IT Department to offer greater levels of uptime.

  • No VMware Site Recovery Manager License is required.
  • The Service Provider/IT Department can have a heterogeneous storage pool that best meets the different consumers I/O requirements and data can be non-disruptively moved across tiers, across arrays, as the I/O requirements increase or decrease.
  • Rather than all of the consumers’ virtual machines running in a single datacentre, they can be distributed across both sites. This enables the service provider to fully utilise the I/O capability of both storage arrays that are providing array resiliency.
  • Failover between the sites is automated using VMware HA, with no manual intervention required reducing again the operational overhead.
  • The loss of a storage array does not result in application failover.
  • Datacentre maintenance does not result in customer downtime.
  • Non-Disruptive data mobility ensures that storage refresh or data movement is a simple task that does not consume operational cycles.

In addition the advanced back-end storage federation opens up new revenue opportunities for offering true continuous application availability for applications such as Oracle RAC.

The VPLEX Benefit

The benefit that VPLEX can bring needs to be considered from both the business and operational perspective.

Business Benefits

clip_image002

Operational Benefits

clip_image004

Summary

The bottom line to this ramble of a thought process is that whilst there are many potential benefits for Service Providers/IT Departments to implement VPLEX between their datacentres the cost model still has to stack up. In order to make this a ‘no brainer’ purchase for a Service Provider/IT Department they really need to be paying very little more than they currently pay for their current replication technologies whilst at the same time enabling them to benefit from the operational savings to reduce the cost of delivering the services.

Time to go speak with the people in charge of our Service Provider and Pricing programmes to make sure we nail this for our all of our potential customers in 2012.