Making Clusters Easy to Use and Easy to Own

Jerry Melnick is currently responsible for defining corporate strategy and operations at SIOS Technology Corp.

Companies have more choices than ever to save money and improve efficiency in their IT environment such as public and private cloud, virtual server environments, and high performance SSD storage. They can even combine a traditional physical server environment with cloud in hybrid configurations for better disaster recovery protection.

However, protecting business critical applications such as SQL, Oracle, SAP, file and print from downtime and disasters in these environments poses a variety of challenges to traditional SAN-based cluster environments. Companies should rethink their approach to high availability (HA) and disaster recovery (DR) and consider alternatives that do not require the use of shared storage.

Options for high availability and disaster protection

Traditionally, companies have provided HA protection by configuring a primary application server and a standby server in a cluster configuration. They use clustering software that monitors the application and in the event of a failure, performs a “failover” of application operation to a standby server. Since both the primary and standby servers use the same shared storage (SAN), the backup server has immediate access to up-to-date data as soon as it takes over operations. This method poses a variety of limitations that may completely offset the benefits of adopting today’s flexible, dynamic data center configurations.

For example, cloud providers do not offer SAN-based clusters for HA protection. Even a fully redundant SAN resides in a single location, making it a potential single point of failure that is vulnerable to threats, ranging from simple power failures to regional disasters. To add disaster protection to an existing SAN environment, you have to purchase an additional SAN that is identical to the first—doubling the hardware cost and adding vendor lock-in. You may find it difficult or even impossible to integrate it with failover clustering.

Clusters without limitations

One alternative is SAN-less HA failover cluster software that eliminates the limitations of shared storage. It works like its traditional counterpart, but instead of using shared storage, it uses block level replication to synchronize local storage in the primary and standby servers. The synchronization process gives the standby server access to a copy of the same data as the primary server.

SAN-less software can be used as a cost-effective add-on to Windows Server Failover Clustering or to provide complete Linux clustering. With it you can use efficient and low cost server side storage. The key benefits of SAN-less software cluster designs include:

SAN-less clusters give you the freedom to configure your data center without giving up HA protection. Use both SAN and SAN-less environments and any combination of physical, virtual, and cloud configurations. Since you don’t need identical hardware at the source and destination there is no vendor lock-in. You can even use your existing hardware.

Leveraging the full benefits

Virtual server environments let you allocate computing resources more efficiently and scale your data center as your company grows. You can create a SAN-less cluster using VMs sitting on any hypervisor using replication to synchronize storage on the primary VM with storage on a standby VM. Locate the standby VM in the same data center, in your DR site, or both. In the event of a disaster, the standby VM can be brought into service with little to no data loss, eliminating the hours needed for restoration from back-up media.

If you install SAN-less clustering software at the Microsoft Hyper-V hypervisor level, you can replicate and failover VMs from one host machine to another. Move data, applications, and entire VMs from one host to another quickly and easily without interrupting service to end users. You can also restore replicated VMs to perform DR testing without disruption to the production site.

Extend clusters for disaster recovery

Extend a SAN or SAN-less cluster to a remote location, the cloud or a node outside your cluster to provide efficient, real-time, block level replication and disaster protection for your business critical applications. If you have a failure in the primary system, the application will failover to the local standby server automatically. If all of your systems fail or if you have an outage of the entire data center, you have a third node in the cloud waiting to take over with a real-time copy of the data.

The software enables failover of applications across geographic locations and cloud availability zones or regions to provide site-wide, local, and regional disaster protection.

SAN-less clusters for high performance storage

You can build a SAN-less cluster with local attached high performance SSD storage to speed application response times and get complete failover protection for a fraction of the cost of a SAN-based cluster.

New configurations are giving you more freedom than ever to configure your data center in more efficient, responsive ways. SAN-less clusters provide the flexibility and ease of use needed to take full advantage of these new configurations.

Industry Perspectives is a content channel at Data Center Knowledge highlighting thought leadership in the data center arena. See our guidelines and submission process for information on participating. View previously published Industry Perspectives in our Knowledge Library.