Following on from our blog last week on how Always On capabilities are a ‘must-have’ to stay ahead of competition, this blog looks at how your business can overcome IT challenges when implementing an Always On solution.
Challenge 1: Removing IT complexities
Today, most vendors require an additional, dedicated, point-solution (HA Gateway) with a different feature set, management tools, monitoring requirements etc. that adds complexity to both operations and automation processes. By implementing a fully integrated solution that combines storage and Always On capabilities this can be simplified.
Challenge 2: Reducing costs
HA gateways are expensive and often priced per capacity. This forces organisations to minimise the number of applications benefiting from Always On, aggravating an already high cost to the business from All Flash Arrays (AFAs) that is about to get even worse. A second cost is added when the HA solution requires Fibre Channel between the sites, taking a toll on the operational budget. The best way to solve this is by encouraging organisations to protect more by not charging for protecting more applications and leveraging cost-effective IP infrastructure that is shared with multiple use cases.
Challenge 3 – Consistent performance, anywhere
Synchronous operations across two storage systems on two sites, increases latency, usually doubling it or even worse. This makes it harder to adopt for latency sensitive applications, regardless of their criticality. In order to minimise the performance impact it’s important to reduce at least a third of the latency on the remote side to allow wider adoption and improve user experience.
With long distance deployments, sending one write to the local storage array and another to the remote one will result in up to double the latency, creating an inconsistent user experience. This is often the result of misconfiguration. It can be easily rectified by simplifying and automating the configuration to enable hosts to differentiate the paths they should use all the time from the ones they should only use only in a failure scenario.
Challenge 4: Variable functionalities
Many solutions that claim ‘Always On’ capabilities vary dramatically in functionality. Some offer a read-only copy on one of the sites and automated failover whilst others have high performance penalties when sending writes to the “secondary” copy. This complicates comparing and deploying these solutions. However, by treating both copies equally, with minimal to no performance penalty between them, hosts can write on both sides (truly active copies) without any failover process.
Challenge 5: Reliability
As any geographically distributed solution, Always On clusters are at risk of a split-brain and require a way to “break the tie” in the event of a communication failure between the two systems.
Optimally, businesses need to deploy a witness, in a third fault domain with redundant networking to each site, to complete a quorum in the event the two systems can’t communicate directly. Making that witness fit the customer’s preferred deployment method (cloud / on-premises) is the vendor’s responsibility.
Meeting challenges without compromise
When Gartner named INFINIDAT a leader in the 2018 GDPA Magic Quadrant one of the key cautions they raised was the lack of a metro/stretched cluster support. We agree!
The InfiniBox Always On Cluster is comprised of two InfiniBox arrays in 2 separate fault domains, and a witness located in a third fault domain, with redundant communications between them. By combining the seven nines (99.99999%) reliability of each InfiniBox array, with Always On clustering, INFINIDAT customers can now benefit from the ‘Always On' Guarantee! This straightforward guarantee allows businesses to design their infrastructure to run Always On, with INFINIDAT providing both the guarantee and the clear remedy (of your choice!)
Want to learn more on how we can help you? Download the ‘Always On’ solution brief or
read about our ‘Always On’ guarantee.