MSP Today Expert Feature
October 07, 2014

High Availability and Disaster Recovery Possible in the Cloud, If Implemented Properly

The cloud is pretty much pervasive in the technology sector at this point, with massive players like Microsoft and IBM (News - Alert) changing business strategies to bolster cloud computing. And organizations of all sizes, from SMBs to Fortune 100 companies, are looking to the cloud for their infrastructure, storage and service delivery needs.

High availability (HA) and disaster recovery (DR) are important components of any cloud strategy, but especially for mission-critical applications and data. Yet, according to Jerry Melnick, COO of SIOS Technology Corp., organizations too often believe myths and hype about HA and DR implementation. And that could prove to be a costly mistake and a major business disruptor.

"While the benefits of the cloud may be clear for applications that can tolerate brief periods of downtime, for mission-critical applications, such as SQL Server, Oracle (News - Alert) and SAP, companies need a strategy for high availability (HA) and disaster recovery (DR) protection," said Melnick. "While traditional SAN-based clusters are not possible in these environments, SANless clusters can provide an easy, cost-efficient alternative."

Melnick is a good source of information on the topic, since SIOS specializes in producing SAN and SANless clustering software. He said there are five popular myths surrounding HA and DR as they relate to cloud deployments and separating the truth from the hype can be very useful as organizations migrate to the cloud.

The first popular misconception is that the cloud is inherently a HA environment, particularly public cloud deployments that offer some measure of data protection through redundancy. However, HA isn’t guaranteed and applications like SQL Server and file servers require additional configuration to automate and manage HA and DR.

Another popular cloud myth is that shared storage is required to protect business critical apps in a cloud with a cluster. A common belief is that Windows Server Failover Clustering (WSFC) may not be used to provide HA in a cloud because it requires a shared storage device like a SAN, and SANs in support of WSFC are not offered in public clouds like Amazon and Microsoft (News - Alert) Azure. Melnick points out that HA for Windows apps is possible in the cloud by adding SANless cluster software configured for a WSFC environment. SANless software synchronizes local storage via real-time, block level replication, which gives apps immediate access to current data should a failover occur.

The third common myth holds that remote replication isn’t needed for DR and that apps and data are protected in the cloud without any additional configuration required. And while massive data centers certainly offer protection against normal hardware failures and unexpected outages, they should still provide protection against regional disasters. Melnick says the easiest way to do this is to configure a multisite cluster within a cloud and add an additional node in an alternate data center or geographic region to extend it.

Many people also mistakenly believe that using the cloud is an all or nothing proposition, when there are a number of options in reality. Customers may use on-premises and cloud data centers in a number of configurations, with on-premises sites using traditional SAN clusters that include a remote cluster member hosted in the cloud.

Finally, many believe that achieving HA in a cloud environment is inherently expensive and complex. In reality, HA cloud clusters may be created easily using SANless clustering software. Users may create a standard WSFC in a cloud quickly and cost effectively since they need not buy enterprise edition versions of Windows apps to get HA and DR.

Edited by Maurice Nagle