Written by Beth Hildreth on February 27, 2015

Moving applications to the cloud can save money for enterprises and provide huge IT and business agility. However, there are challenges that IT managers need to overcome to meet the High Availability (HA) requirements of business critical applications in cloud environments.

Failure is not an option for enterprises that want to have business continuity at all times. Availability goals can be achieved in a variety of configurations depending on the types and condition of environments on which the applications run.

The traditional way

Traditionally, the preferred way of achieving HA for critical applications is shared storage clustering. This is ideal for applications operating on a SQL server habitat using Windows Server Failover Clustering (WSFC) for configuration.

In this scenario, IT managers typically cluster together two or more servers sharing the same physical storage set up through a Storage Area Network (SAN). One server supports the critical application, but in case of failure, the clustering program moves the application to another server in the cluster. Even after a failure, the application can continue to operate without the loss of data because all the cluster servers share the same server.

The SANless way

In environments where traditional shared storage clusters are not possible, SANless clusters are a fitting alternative for HA and Disaster Recovery (DR) protection. SANless clustering allows IT managers to group servers together without the need for shared storage. It uses efficient replication to synchronize local storage on each cluster link, so that the setup appears as shared storage to the WSFC. It can be used in cloud, hybrid cloud, virtual server, and physical server environments.

SANless clusters provide high availability of applications and data. Enterprises can also take advantage of cost savings, flexibility, and efficiency by using virtual servers and cloud applications. The SANless method also eliminates downtime and data loss and helps IT departments achieve high availability goals.

Database mirroring

Database mirroring is a method of maintaining two copies of a single database that reside in different servers at different locations. Called the principal server and the mirror server, these two servers communicate and cooperate as partners during database mirroring sessions. It is important that the principal server performs its redoing function on every activity it does to the mirror server in the original sequence as quickly as possible to eliminate data loss in case of unexpected failures.

Providers of high availability solutions are not lacking in options that customers can choose from, such as replication, transaction log shipping, virtualization, geo-clustering, and load balancing. No matter how sophisticated they may be, HA and DR solutions cannot serve their purposes to the maximum if IT managers do not observe basic HA best practices. They likewise need to meet the proper HA architecture, engineering, and management prerequisites that their environments require.