Stupidly Obvious AWS Pitfalls That Trip Up Cloud-Hungry Businesses

Stupidly Obvious AWS Pitfalls That Trip Up Cloud-Hungry Businesses

Migrating a business onto Amazon Web Services (AWS) or another cloud provider can be a highly complicated process. Then there are the blatantly obvious dos and don’ts that nevertheless trip up an embarrassing number of companies.

Storm picture from Shutterstock

We spoke to Mark Randall, the chief customer officer of AWS’s Australian premier consulting partner Bulletproof. He shared some of the most common pitfalls that cause his customers unnecessary headaches while migrating to and operating in a cloud environment. Most of these might seem pretty obvious, but they nevertheless cause businesses to lose money and waste resources, time and time again.

Treating cloud like traditional IT

“We often see customers approaching cloud and AWS with the same mentality and mindset of in house IT or traditional hosted environments. They’re not really factoring in that the architecture is fundamentally different.

“For instance, in traditional IT you typically invest in very expensive pieces of hardware on the basis that they’re going to give you a fixed amount of uptime, whereas the approach of AWS Is to design your application on the basis that failure is going to happen and design it so that it can continue to run in that environment. We see a lot of customers who don’t deploy in two separate availability zones which means they’re taking a big risk when it comes to experiencing unnecessary downtime.

“Another example of customers approaching cloud from a traditional IT perspective is that you can’t just take the applications you have and expect them to work in the same way. A lot of traditional applications are not built for cloud; whether it’s a licensing consideration or the fact that the application is not loosely coupled; it doesn’t scale out horizontally, it only scales up. So you have to make sure you really understand the architecture and how to get your legacy apps that weren’t built for cloud working effectively in that environment.”

Failing to keep your workforce appropriately briefed

“Another thing we see is inadequate internal communication or external communication. The textbook example is when the marketing department plans some huge campaign or the CEO is being interviewed on Today Tonight and nobody tells the IT team or the external cloud provider. As a result, they find out that they aren’t set up with the required capacity when it’s too late.

“It’s not a technology problem, it’s just that the IT partner or team wasn’t expecting it because they weren’t informed.”

Not factoring in the real-world cost of downtime

“One thing we’ve noticed is that customers don’t always factor in the cost of downtime. This is obviously an important consideration if you work in e-commerce or run any type of transactional website. When people look at the cost of Option A and Option B as it relates to infrastructure they don’t necessarily relate that back to their business and what effect downtime can have.

“For example, if you have a mission critical e-store that’s doing $5 million in revenue a year, every hour of downtime is costing you seven grand. So when you’re making that comparison between a solution that gives you 99 per cent uptime or 99.9, it’s really important that you work back and calculate what that means for your business. 99 percent uptime is over seven hours of downtime a month while 99.9 is 43 minutes of downtime per month.

“You need to ask yourself: what is the value of the business I’m supporting on this infrastructure? How much uptime am I expected to get on the investment that I make? This is something that very few customers actually do. Instead, they go for the best price or the best features. If your site is carrying thousands of dollars of business per hour, the cost of downtime should be an important part of the decision.”

Neglecting security

“One of the problems with cloud and the fact that it makes everything self-service is that sometimes it’s easy for people to forget it’s a complex technology requiring specific skills if you want to do it properly. You still need to understand the principles of security and networking or at least work with a partner that does.”

Failing to use the technology at your disposal

Cloudcheckr is an AWS Technology partner that conducts regular surveys and audits of Amazon customers to find out how effectively those customers are using the services. In one of the recent reports, they identified that 98 per cent of AWS users have issues maintaining uptime and 91 per cent were failing to do storage volume snapshots on a timely basis. Users also repeatedly failed to optimally configure Auto Scaling and ELB.

“All of these are issues that with proper configuration and proper management and support can be avoided.”