Can we increase availability and reduce any downtime? Perhaps this is an issue facing us right now. After all, we need IT infrastructure that we can count on when facing such problems as network outages, equipment failure or power issues. There are two solutions that address both, the Fault Tolerance (FT) strategy and Clustering. Let’s see the importance of both and finish with what I consider to be the best solution.
Why is Fault Tolerance important?
An FT system is akin to High Availability (HA) but goes one step further by guaranteeing zero downtime as opposed to the four nines of availability. If you have a system that cannot be allowed to fail, FT is the solution.
Stratus ftServer hardware has focused on helping businesses (where downtime isn’t an option) for some time now. Whilst enabling some of the highest levels of uptime in the industry, each ftServer is scalable and can be tailormade to suit. Along with that, there is an additional fault-tolerant Stratus storage array – ftScalable – which can provide an incredible 182Tb of fault-tolerant usable disk space.
The requirement to have inessential systems both consuming resources (RAM, CPU) means the need to have 100% extra resource requirements resulting in performance degradation. Successful FT systems hinge on specialised hardware to detect faults so they can quickly switch to a redundant component. Although this switchover seems flawless, the redundant components stop processing and a high premium is paid regarding hardware cost and performance. What is the most common reason for downtime? Yes, it’s software failure and the FT model does not address this. When systems are not working to the standard expected because of component failures, it can become harder to diagnose those issues. This means that more resources are needed to evaluate the health of a system designed for FT (expense).
Clustering with SIOS technology – the difference
An HA solution such as Failover Clustering cannot deliver zero downtime for unplanned hardware failures. However, it does protect against a wider range of failures up and down the stack and even provides application health monitoring, reducing cost in the long run.
SIOS technology allows you to build clusters the way you want, using a choice of server and storage hardware to protect business-critical applications in both Windows and Linux environments. They can make use of any combination of physical, virtual or cloud environments.
SIOS clusters adopt a HA configuration which supports synchronous and asynchronous replication across multiple servers and data centres. It can use high-speed storage solutions such as PCIe-based flash and SSDs (Solid State Disk). The performance is fast and much cheaper. Not only are SQL databases protected, but they will run very smoothly…no performance issues. Although FT seems the logical solution to having no downtime at all, there is no protection from software failure because both VMs can come crashing down. Failover Clustering will failover and recover on another independent system.
So, in short, Clustering wins the battle for me as it’s the ‘lesser of two evils’ – yes, experience a little downtime but have no performance degradation. Fault Tolerance, performance degradation with the issue of cost but with no downtime.