These days, applications and data are the lifeblood of most organizations, and everyone expects that the applications and data they need to get their work done will be available. But ensuring true application and data high availability (HA) – by which we mean the assurance that you can interact with them at least 99.99% of the time – can sound like a costly proposition. A clear understanding of where to apply HA and how to do so cost effectively can more than justify the cost. In fact, it can protect you from the extensive costs and consequences of downtime and disaster. The question is, how can you determine how best to invest in a high availability infrastructure?
Weighing high availability value for money
The place to start is with a clear understanding of the value of different applications to your organization: Which applications, or combination of applications, would cause the greatest financial loss to your organization if they were to go offline for more than — let’s be specific, here — 4.38 minutes in a given month? That’s the maximum amount of unplanned application downtime that would be tolerated in an HA environment.
For some companies that might be a customer-facing order entry system or an application that performs high-value transactions in real time. If the loss of revenue caused by one of these systems being inaccessible is greater than the cost of a high availability infrastructure, it becomes pretty easy to justify supporting those applications with an HA infrastructure. If the application suddenly goes offline — for whatever reason — your primary infrastructure would fail over to secondary infrastructure, and your applications could be back up and running in seconds, ensuring continued support for your customers and business operations.
But don’t stop at potential revenue loss
The financial significance of application downtime is not the only factor to consider. There may be reputational costs that could be just as injurious. If your company brand is all about reliability or security, for example, it’s not a good look if your customer-facing systems go offline and you can’t bring them back for several hours. Why would customers trust you if you can’t keep your own applications running? Again, the ability to bring your customer-facing systems back online within moments may be readily cost-justified if the reputational costs of being offline would ultimately prove even more damaging to your business.
What about the high availability of other applications?
In identifying your critical applications, you will quickly discover that there are far more applications that, while indispensable to your organization’s day-to-day operations, are not so important that your business will suffer catastrophically if they suddenly become unable for more than a few minutes. Their being offline will be annoying, but it may be difficult to justify an investment in a high availability infrastructure because individuals are annoyed.
Still, there are things that you can do to ensure that these not-quite-critical applications remain as accessible as possible without the full protection of an HA infrastructure. Unsurprisingly, some of the best things you can do are the things most often ignored:
- Have a backup plan. If your storage system suddenly fails and you don’t have an HA infrastructure with an up-to-date replica of the data in that storage system, at least you can restore most, if not all, of your data from backups. It might take a few hours, but that’s better than having no backup at all.
- Practice for outages. Restoring operations after a failure often requires cooperation from several IT departments – network, storage, OS, and application. Practicing for outages will help your personnel identify vulnerabilities and ensure that they know what steps to take if an outage actually occurs.
- Patch your systems. Every operating system and application patch that comes out is intended to lessen the possibility of a system failure or security compromise. The longer you put off updating your applications, the more vulnerable they will be to factors that can compromise your access to them.
- Monitor your infrastructure. If you’re encountering hardware or software faults, address them as quickly as possible. Don’t wait till they cause your applications to crash. Consider your infrastructure loads as well; it may be that an investment in load balancers would facilitate the distribution of system loads and reduce your risk of system failure considerably.
- Monitor and manage infrastructure capacity. If your storage infrastructure is filled with unused applications and data, do some housecleaning. There’s no value in fragmenting your critical applications and databases because your storage is filled with old data and application stores that no one is using.
These are some of the most cost-effective ways to increase application availability without investing in an HA infrastructure. They’ll also lessen the risk of reputational damage if your applications go offline unexpectedly. What, you didn’t even have a backup plan? You never bothered to patch your systems? Your users may be annoyed, but you don’t want them posting comments on Facebook or LinkedIn.
Eddie Smith is a Solutions Architect at SIOS Technology.