Cloud Computing is Not Easy

There, said it! – Well, it’s not. Anyone who has looked at migrating a service, product or environment to the cloud soon realises that the marketing does not match the work that has to be done. Even if you are setting up a new business or infrastructure some things are more complicated than initially seem.

It’s all well and good stating that, but why do we think this is the case? Just a few items that sprung to mind during a recent project that are general items that should be considered.

It does what?

What is quite eye-opening about migrating to the cloud is that a lot of organisations may not have any idea of how much their current environment does.

This could be for lots of reasons, and most likely not from a complacency – more the typical expansion of services over time and the pressure for all teams to document it, people moving on, taking over from “legacy services” and so forth – even the dreaded shadow – IT…

We’re not just talking about the number of VMs, cores, memory and so forth but also quite “simplistic” items such as

• What services are currently running and what they bring to the organisation
• Where those services are located and how they are accessed
• How much data is being processed over the network or internet
• How the users connect and work on systems (be it via SaaS, VPN, RDP, TLS and so forth)

Some organisations may not even realise they have an application running until they ask users what they use every day. It happens.

Point 1 – Know what you have, what it does and how it is used. 

Need help? You could utilise an advisory service to do so. 

It costs what?!

Of course you are looking at the costs right? That’s most likely one of the key motives to move to the Cloud.

Humans being humans, we always try the path of least resistance – in this case take whatever is “there” and place it “here”. Well, that’s great logically, however replicating what you might have with one service / on-premise items can cause a bit of shock to the system if it’s not reviewed.

We’re not just talking the cost, factor in how your users will connect to it, what the impact may be on business processes and procedures or even changes in software / software versions may bring – leading the project from a pure cost stance could actually make things worse.

The SaaS version of your on-premise solution, for example, may not have the same capabilities or is more rigid in the way it works.

Items to watch out for:

• Moving to SaaS version of software? How does that impact your processes?
• Incoming and outgoing data charges (you might pay for a flat fee at the moment)
• Storage charges (got a server with oodles of space? Need it?)
• Replication charges – backing it up to another zone or re-creating a DR site? It will cost.
• Those users connecting in? Adds up.
• Have a need to connect to on-premise items or another location? Add that on as well.
• Are you ready for the change to OPEX? Can your finance team facilitate monthly fees?

Food for thought. You could consider putting the scalable stuff in the cloud and perhaps the heavy lifting elsewhere? Does it all need to be in a single location? It may not be cheaper with the same supplier or your existing solution is just fine for some bits and can be re-purposed?

Point 2 – Check the costs and optimise. CAPEX can mainly become OPEX and your finance team may need to be involved. 

Worth pointing out here that Nine23 provide a monthly flat fee for all services to ease that complexity and worry. 

It’s stored… where?!

It’s in the cloud right? All taken care of? No need to worry about all that security faff?

Even with a SaaS solution, it’s worth due diligence to ensure your data is safe. How they handle it, who can access it, how they protect it, secure it in transit and ultimately delete it when it needs to be are all perfectly valid questions.

Data sovereignty can also be important depending on what the sensitivity of your data requirements are. Needless to say, if it’s personally identifiable information you’ll be wanting to call on your GDPR talent to verify that all is well – thinking about it, you should be regardless of where you are placing your services.

In some Cloud providers T&Cs there is nothing stopping them moving your data from your chosen location at will. It’s not to say they will do it; you just will have signed to say it’s OK for them to anyway and to anywhere.

American based providers may also be subject to the PATRIOT act – another thing to check.

Again, your data needs may vary. Inventory of office biscuit stocks and tea levels are vastly different needs to processing legal documentation.

Most providers now have a secure by default mentality – EXCELLENT – hopefully gone are the days where someone leaves an S3 bucket open to everyone, but probably best to get someone to check it all over before, during and at regular times after it’s all up and running.

Point 3 – Read the small print. Check out your legal commitments, never hurts to double check. 

It’s due… when?!

If you read the material around migrating to the cloud – it makes sense, you take your systems, you press a few buttons, change a few things – boom! – you’ve moved to the cloud.

Sadly, it’s not quite there yet. Just a few observations:

You’ve likely got to move a ton of data, maybe some legacy VMs, sort out how it’s all going to fit together, finding the right services, getting the networking up and running, how does it impact the users, how do you cut over to the new system, do an alpha or beta test. Are you re-architecting your solution to be cloud native? – great! but that takes time as well.

Factor all of these things in and the business processes you may need to change, and with all of the above – it’s going to take a while to do it properly.

Point 4 – It takes longer than you think.

Written by Richard Perry – Nine23’s Infrastructure Engineer