Ok so you can’t actually setup a failover cluster in the traditional sense on Amazon Web Services (AWS) EC2 instances (virtual machines). Essentially AWS does not provide a disk storage option that allows you to provision cluster aware, shareable disk devices.

However, all is not lost, SQL Server Magazine’s Denny Cherry article (here) highlights a neat solution. The SteelEye Datakeeper Cluster Edition (website link here) provides real-time data replication that integrates with Microsoft Windows Server Failover Clustering to enable multi-site clustering while eliminating the single point of failure that a traditional shared physical disk cluster resource represents.

Standout features of SteelEye DataKeeper include:

• Provides continuous real-time host-based block-level replication of data insuring multiple copies of your data is available at all times
• Replicate data while running within guest virtual machines or replicate entire virtual machine files while running within the Windows host server (Hyper-V, VMware Server or Microsoft Virtual Server)
• Multi-site clustering delivers the highest RPO and RTO available. DataKeeper enables this technology without a large investment in array based replication.
• Optimized for WAN replication with 9 Levels of Compression, WAN acceleration to provide maximum throughput for high speed networks and Bandwidth Throttling
• Replication works across disparate storage arrays, and with no regard to make or model of server or storage, allowing the optimal use of your available resources

Indicative cost is US$4995 per server, the cost doesn’t seem too bad, and especially considering that you only need SQL Server Standard Edition and not the pricey Enterprise Edition that is typically required for Always On Availability Group implementations.

Here’s a link to a YouTube video (here) demonstrating the Datakeeper clustering feature using a Windows cluster with SQL Server.