High Availability Is the New Black

Chic, slimming, and goes with everything

A little over a month ago, someone sent me an article about Target experiencing downtime for a couple of hours at its checkout lanes, and I have been dwelling on it the last few weeks. While instances of technical downtime in business are nothing new, the fact that it affected so many stores worldwide had me interested in how edge computing might help similar organizations avoid downtime in the future.

In today’s technology-driven world, IT downtime can cause some serious problems. While this particular incident at Target undoubtedly caused some loss in sales, these types of incidents also cause loss of reputation as customers don’t receive the service they expect.  A well-established company like Target can withstand a couple of hours of downtime and long checkout lines every so often, but smaller businesses might be so adversely impacted that they never recover. 

So, back to my thought process, how could edge computing make a difference? Edge computing is about putting the computing resources close to where they are needed. When there are devices at branch locations like point-of-sale cash registers at retail stores that are all connected to a centralized data center, an outage at that central data center can affect all the branch locations. That is what we call a single point of failure.

By putting an edge computing platform at each branch location, a failure at the central data center does not need to bring down everything. Each branch can run independently. A solid virtualized environment can run all of the different applications needed to provide customers with the high-tech services they have come to expect. 

You might be thinking “Why hasn’t this been done before?” or “On-prem infrastructure like this is cost-prohibitive” or even “What about the cloud?” Until very recently, it was cost-prohibitive to implement the kind of infrastructure that would make this work: highly-available infrastructure. Creating a highly-available virtual infrastructure involved a sizeable investment in a shared-storage appliance, multiple host servers, virtual hypervisor licensing, and then a disaster recovery solution. As for the cloud, well, that is a solution in theory. In reality, and as you have seen in the news, even cloud infrastructures fail, and it becomes an even bigger single point of failure for everyone running on it. Cloud also poses other challenges that edge computing can solve such as latency, security, autonomy, and more

Fortunately, hyperconverged infrastructure (HCI) has consolidated those components into an easy-to-deploy, low-cost solution. Still, not all HCI solutions are cost-effective for edge computing. Some HCI solutions are still designed like traditional virtualization architectures and emulate SAN technology to support that legacy architecture. This storage emulation results in resource inefficiency requiring bigger systems that are not cost-compatible with edge computing.

HCI with hypervisor-embedded storage like Scale Computing provides can offer smaller, cost-effective, highly-available infrastructure that can allow each branch location to run independently even if the central data center goes down. A small cluster of three HCI appliances can continue running despite drive failures or even the failure of an entire appliance. There is no way to prevent downtime completely, but edge computing with the right highly-available infrastructure can insulate branches to continue operating independently. 

The central data center is still a vital piece of the overall IT infrastructure because it consolidates data from all of the branch locations for analysis to make key business decisions. That doesn’t need to change with edge computing. On-site edge computing platforms can provide local computing while communicating key data back to the central data center for reporting and analytics. By taking the single point of failure out of the central data center, outages at any location need not have far-reaching effects across the whole organization.

As we see more and more technologies become commonplace in all aspects of business, industry, and our daily lives, high availability is going to be more and more important. HCI is quickly replacing traditional virtualization infrastructure and making it more accessible. Whether your organization is one small location or has thousands of small branch locations, you’ll want to consider HCI for highly available edge computing.