1. Configurations and ‘always on’
There’s a longstanding principle in data analytics that ‘garbage in equals garbage out’. The same principle applies to cloud migrations: shifting a garbage system somewhere else doesn’t improve it. Consider this the first task. Ensure your servers, software, storage are appropriately structured, configured and priced – before any ‘lift and shift’ activity.
There’s an emerging term for this discipline: FinOps. Short for financial operations, this promotes shared responsibility for cloud computing infrastructure and costs. Too many firms move into the cloud expecting a cost saving. They don’t realise the cloud runs 24/7, even if their business shuts at 5pm. Poor configuration might mean paying for; 24/7 availability or unneeded features. These run up serious costs for zero benefit.
It happens easily too. When you run IT on premises, using every resource (whether needed or not) doesn’t matter. You’ve already paid for it, so why not. In the cloud, things differ. Every feature, every resource, every minute… comes at a cost. It’s a different discipline to the on-prem world.
2. Going serverless isn’t always a great idea
This relates back to point 1, configuration. But ‘serverless’ is a specific problem routinely cropping up. Serverless computing seems a great idea, but it has limitations as many firms have discovered.
While it can drive down costs as the big cloud providers allocate compute / other resources on demand, there’s a performance penalty. This has resulted in things going full circle, and a realisation that maybe the old ways were best after all. This means configuring and paying for ‘traditional’, albeit virtual servers in the cloud. Reverting to this configuration often blows initial financial calculations out of the water. It quickly turns a cut price cloud solution into an expensive opex commitment. Which brings us to…
3. All that opex adds up fast
This is one of the most frequent cost blowout scenarios. Sure, an opex line item doesn’t hurt as much as a big capex hit. But it comes once a month, and those month-ends just keep coming. Cumulatively, the opex can result in a lot more than the capex.
Then there’s ‘the Netflix effect’ and the inevitabilities of shadow IT. The Netflix effect is when it’s just 15 bucks a month for one service. Before you know it, there’s YouTube Premium, Disney+, Amazon Prime, Neon, Hulu, HBO Max and a VPN thrown in for good measure. Your services have multiplied. This happens in an enterprise as readily as it happens in a household.
Furthermore, the cloud is made for shadow IT. Applications and resources are easily accessed and purchased (and can include anything from Zoho to Salesforce.com). It’s just as easily forgotten while the company keeps paying and paying. Never mind the app used for a once off project has become shelfware, along with other misc. services and storage utilities. Aside from security concerns, these indiscretions quickly fatten that opex line for no discernable benefit.
The cloud is amazing. Especially when done right
The cloud’s promised cost savings can and do result, but only when it’s done right. And it’s routinely reviewed.
This is why CodeBlue spends time understanding your business’s entire technology and operational stack. We see how IT supports operations today, and help plan how IT supports organisational goals tomorrow. We then formulate a roadmap that looks ahead three years. And finally, reassess IT and business conditions as they change over time.
We often advise not lifting or shifting anything within the first year. Only after a thorough understanding of your situation can migrations be made – if necessary. The cloud can deliver cost savings, extra flexibility and easier access. But only when it’s done right. Just like everything else, discipline, rigour and insight are required for success.
