Could a cloud migration disrupt your operation? Absolutely!

Steven Spielberg’s 1975 thriller Jaws is widely accepted to be one of the greatest films ever made, establishing the ‘summer blockbuster’ business model that continues to this day.

Jaws hauled in more than $470m at the box office and won multiple awards for both its music and editing. In 2001, it was selected by the Library of Congress for preservation in the United States National Film Registry for being “culturally, historically or aesthetically significant”.

You might be wondering what Jaws has to do with cloud computing. Well, the film may have broken box-office records, but things could have easily been very different.

At the time, Spielberg was criticised for being inexperienced and for his insistence on shooting the picture on the ocean – the first major film to do so. Principle photography was scheduled for 55 days but ended up taking 159; the original budget of around $3.5m almost trebled to $9m.

The delays and overspend were largely as a result of the three mechanical sharks built for the film, at a cost of $150,000 each, all three of which kept breaking down.

The sharks had only been tested in fresh water; the first time one arrived on location and entered the ocean, it promptly sank to the bottom. Within a week, the saltwater had corroded the electronics, the internal frames and the painted skin.

All these issues could have been identified and addressed far earlier in the production schedule by conducting the appropriate testing and due diligence. Exactly the same thing applies to your cloud migration.

Fail to prepare, prepare to fail

“The key to a successful migration is planning and stringent testing supported by very comprehensive back-out strategies. You can’t over plan these events, particularly when it comes to mission critical – ‘heart and lungs’ – applications,” says Paul Norton, solution director at cloud specialist Rackspace.

Paul has seen first-hand what can happen when that isn’t the case.

An aerospace business sought to re-purpose its on-premise data centre, a goal that required moving several systems into the cloud, including a mission critical, bespoke manufacturing application.

Appropriate due diligence was conducted, and the application was run through its paces once it was built out in the target cloud environment. However, it wasn’t tested at load.

“Testing validated whether the application could be installed, whether the data integrity was correct, whether the security compliance was intact, whether it was operational and in-line with expectations. But this was, in effect, only a single user test,” explains Paul.

The technical challenges associated with moving large volumes of data over long distances reliably, means WAN links tend to have lower bandwidth and higher latency than LAN links.

The application in question wasn’t optimised to be run off-premise via a WAN link, and within five minutes of it receiving full production traffic, the network became congested and production stopped. As a result, disaster recovery (DR) was invoked and a full back-out plan came into effect.

“This application was vital, operating without it simply wasn’t an option,” Paul continues. “The time needed to reengineer the application wasn’t available; so, in this case the remediation strategy was an unforecast spend of £1m to quickly deploy a server farm in order to simulate a LAN environment and control how the application behaved.”

Unexpected costs

The aerospace business imposed on itself a date to exit its data centre which, perhaps, unnecessarily compressed the project timeline. That meant that the level of due diligence and testing required wasn’t possible.

Though the end result achieved the desired outcome of exiting and repurposing the date centre, no management team wants to be hit with an unexpected bill for £1m. Thankfully, such a bill is a rarity.

That being said, rarely is a major IT project completed 100% on time, within budget, with no unexpected bumps in the road. That’s where the need for thorough planning and comprehensive contingency strategies come to bear.

“Upfront planning is critical; if you plan for things in advance then you have the time to deal with them should they arise. It’s always better to have that contingency factored in and deliver something earlier rather than later as that’s when additional costs can quickly start mounting up,” says Paul.

Paul also advises businesses to adopt a phased approach to cloud migration, not an all or nothing, overnight transformation.

“Few businesses have a mature, well-thought out and structured cloud strategy which tackles all the success factors we’ve spoken about previously. I would advise starting with a multi-cloud, hybrid strategy, which means thinking earlier in your journey about containerisation [encapsulating an application with its own operating environment] and Infrastructure-as-a-Service [computer infrastructure delivered on an outsourced basis].”

The rationale for that approach, according to Paul, is that it allows businesses to learn what their operating environments are, whether their enterprise applications are good candidates for cloud migration, and what dependencies they have on particular systems and applications.

“That understanding will help inform whether you need a greater bias towards public or private cloud, as well as where you need to invest in terms of new app development.”

Cloud – right for everyone, not for everything

Regardless of what form of cloud your organisation chooses to embrace, you need to know which applications you should migrate, and which should remain where they are. That knowledge only comes from defining the business outcomes you want to achieve and fully understanding what’s required to gain them.

Many companies, for example, successfully complete a cloud migration, but only realise half the efficiencies they expected. That can be because they haven’t invested in the appropriate automation, standardisation and tooling required to deliver operational efficiencies.

The word ‘journey’ may be over-used, but the principle remains – as does ‘starting small’ to gain valuable experience.

“This isn’t about creating a big bang, it’s about pragmatically and proportionally adopting and changing your organisation to a point where the business benefits you projected are being realised,” Paul concludes.

“Identify how and where your business benefits are going to be derived from and make sure that you’re investing in the right places to start with. Cloud is right for everyone, but it’s not right for everything.

“Cloud can bring tremendous benefits and opportunities if you understand how, rather than seeing cloud as the solution for all your issues.”

*All images courtesy of Depositphotos