For many years, Windows Server Failover Clustering (WSFC) has been used to provide HA and DR protection for SQL applications in traditional onsite environments. Until recently, managing site failures has been very complex and expensive, requiring large investments in specialized hardware and software as well as the availability of a second data center site.
The cloud offers an extremely attractive and cost-effective second site to locate a cluster member and use in handling failover when the local site fails. However, because clusters require a shared SAN between all cluster members across local and cloud resources, this has not been a practical alternative.
In these cases, SANless software can be used to provide high availability across local and cloud resources.
Companies are moving more business operations to cloud and hybrid cloud configurations. Others are adding cloud failover instances to their physical server environments for disaster recovery without the cost of building out a disaster recovery site. Yet, misperceptions and concerns about high availability and disaster protection have made decision making difficult for companies moving their mission-critical applications (i.e., SQL Server, Oracle, and SAP) to the cloud, where traditional shared-storage clusters are impractical or impossible. To fully leverage the benefits of cloud and hybrid cloud environments, companies need to separate myths from facts and consider a new approach to HA and DR.
In this slideshow, Jerry Melnick, president and CEO of SIOS Technology, takes a closer look at five common misconceptions surrounding high availability in the cloud and how SANless clusters can help overcome some of the obstacles.
Public cloud deployments are automatically high-availability environments where application downtime is negligible. This is not true. Clouds are not high-availability environments unless you add HA protection.
You cannot protect business-critical applications in a public or private cloud by using a cluster. The truth is, HA is a physical deployment with a failover cluster using shared storage (SAN). Public clouds (e.g., Amazon EC2, Azure) have no concept of a cluster-aware shared storage.
Applications and data are protected from disasters in the cloud without additional configuration. The reality is that cloud providers experience downtime and regional disasters like any other organization.
You can either run your application in the cloud or in your on-premise data center, but not both. In truth, you can have both. You can use your on-premise data center as your primary site and the cloud as your hot standby DR site. On-premise servers can be SAN-based or SANless clusters or a single server not participating in a cluster.
Creating an HA environment in a cloud requires complex scripting, specialized skills or added complexity. The opposite is actually true. HA cluster in a cloud is very easy without adding another SAN with SANless clusters.