How to manage costs and consumption in a cloud services environment

Posted 1 August 2016 by David Spratt

We talk to dozens of small and large companies around the country to help them modernise and streamline their ICT needs. We frequently hear about an issue that’s important to managers everywhere: a supposed lack of control over the growth of these as-needed costs. In many cases, the concern isn’t a technical issue, it’s a management one. It is also one that can be modelled and controlled.

You can always spend more money on cloud services, but you shouldn’t be surprised by your monthly bill. In an ICT/cloud infrastructure model there are three basic levers that need to be managed in order to control costs and consumption. Let’s take a look at each of them.

ManagingProjectScope

1. Establishing proper authority to buy services

Picture a junior technical person happily sitting at their desk with ideas running freely through their inexperienced but giant brain. “Eureka!” they cry, “I can solve my technical problem by using loads of storage on a huge server”.

Reaching for the department’s username and password, our intrepid techo accesses a public cloud service provider, nails up the required infrastructure and sets to work.

Our junior techo never actually switched off the environment they created; in fact it was so useful that it became the test bed for the whole department. Nor did they even have the delegated financial authority to spend the money.

Two months later a finance person, waving an eye-wateringly large invoice, arrives in the IT manager’s office. The company was committed and the money spent before you could say “breach of process”.

2. Controlling storage and backup growth

For the past 20 years organisations have grown their storage at around 70 per cent each year on average, as shown in the diagram below.

Data Table

The diagram shows you were able to negotiate a 20 per cent price reduction every year for six years while backing up 70 per cent more data annually. Monthly backup charges have grown from $200 per month to a whopping $930, despite aggressive price reductions.

3. Monitoring the proliferation of server infrastructure

Our intrepid junior techo has moved to an organisation with more “flexible” IT policies (that means no policies, in case you needed clarification). Techo’s new boss has just turned up waving an unexpectedly large invoice and demanding a reduction in the number of servers.

We can’t do that! says our techo, These servers are all essential to the running of the business.

Flummoxed, the IT manager storms off, wondering just how to handle the inevitable conversation with Ms Money, the chief financial officer. Ringing in his ears are the words: “Don’t give me problems – I want solutions”.

How to manage each cost contributor

Authority to buy services

A good thing about most cloud service providers is they offer some form of budget control solution, however, organisations need to understand the spending risk and put in place internal policies around just how and when cloud services are consumed.

Storage and backups

I have yet to find a multinational backup software provider that seeks to reduce the amount of data their users back up – hardly surprising as they charge by the Gigabyte.

To get on top of this problem, measure performance against agreed limits on storage and backups, and ensure technical people apply tools like deduplication (ie, only backing up one cat video shared on email), aging non-critical data, and long term archiving.

Measuring performance requires active monitoring by those in senior positions. I use a tool that reports in real time. and used properly incentivises the IT team to actively manage the challenge, and delivers the satisfaction of real financial results for their efforts.

Server infrastructure

Just as with backup software suppliers, server support providers are hardly incentivised to shut down servers when they charge a fee for every server they manage.

The answer for managing costs is the same as for storage and backups: agree overall targets, and drive to these.

Remember, although some servers are actually mission critical, not all are. Switching off a system just to hit an arbitrary target could be catastrophic, so key stakeholder input is as important when switching them off, as it is when switching them on. Monitoring server utilisation can quickly identify significant savings opportunities, and places the onus on the technical folk to justify why a server can’t be switched off in the low season.

Baseline what you have, analyse the projected and actual costs, and monitor the inevitable ebbs and flows of a dynamic IT cloud infrastructure. The rewards will be huge, the savings real and the risk negligible when you get this right.

If you’d like help modelling your ICT management, get in touch with me using the details below.

Tags:

 

View Articles

How can we help?

Contact us