• silicon.com
• Technology
• Software

By silicon.com, 8 April 2009 11:25

COMMENT

silicon.com's panel of experts explains what factors to consider before you move any services into the cloud.

Q: I want to move the majority of my organisation's business-critical services into the cloud. Is this a good idea?

Clive Longbottom, Quocirca analyst:
It depends - the key is to create a triangle with the various processes that your organisation uses. At the bottom, there will be commodity processes - those that the vast majority of organisations have to do, things like booking vacation, ordering stationery, payroll, expenses and so on. These are services that are pretty well available and proven in the cloud already, and keeping them in-house is just costing resources in the form of hardware, power, maintenance and humans running around keeping it all working.

Above these processes are the differentiated processes - those that are similar to other businesses operating in your vertical market, or of your size, or in your location. Processes such as legal governance, high-value supply chain and so on. Again, you may find that there are specialised cloud providers in these spaces that can do the job better than you can internally, and even if not, breaking these processes down into representative tasks can identify where external services can make the processes more effective.

This then leaves you with the unique processes - those that make you the company that you are. Keep these in-house, invest strongly in them and innovate as much as you can. By putting commodity and differentiated processes in the cloud, there should be resources and money freed up to do this - and surely that's good for everyone.

Naked CIO, IT leader and silicon.com columnist:
As a concept, cloud computing is a great idea but in practice whether to use it depends a lot on the organisation.

The effectiveness of cloud computing is dependent on current infrastructure, organisation size, risk, cost, application and hardware integration requirements among many other factors. Cloud computing has to be part of a larger enterprise infrastructure strategy - do you have one? If not spend your time on deriving this first before you stare at the cloud.

Also extremely important is to consider business strategy. Cloud environments support mobile and agile working environments very well but also create massive exit strategy issues should you decide in 12 months to manage your own environments again.

There are variations in the definition of cloud specifically in whether you operate your own cloud or essentially outsource (the Google or Amazon model). Both of these have upsides but can also provide logistical and cost complications that could compromise the business case.

Yes cloud is a good idea for the right environment, strategy and business model. Make sure these are in line and present the best strategy that supports these objectives.

Peter Cochrane, tech guru and silicon.com blogger:
The first question to ask is: what kind of business are you in, and what is the IT capability of the people?

The second question is: what kind of equipment have you invested in, and how many people do you have working out of the office?

The third is: are you thinking of asking people to use their own laptops and desktop facilities?

The fourth: do you intend to build your own cloud, or are you going to outsource?

These will dictate the kind of cloud you should go for and your future operational mode.

Remember that you are trying to afford new degrees of freedom and flexibility in order to up productivity while maintaining the integrity operations and security of access and data.

Overall this is about efficiency and cost saving but these should not be at the risk of operational fidelity.

I favour migrating technology and workforce slowly from where you are to where you would like to be - learning and adjusting as you go.

So let's now tabulate the pros and cons.

Pros:

• Reduced hardware, software, network, support, accommodation and environmental costs
• Greater flexibility, mobility, utility, operational stability, process control and activity monitoring
• Faster updates, procedural changes, training and debugging
• Better security, scalability, use of resources, control of data access and recovery from catastrophic failures

Cons:

• Dependence on centralised server and services that are in-sourced or outsourced, plus the need for fixed and mobile bandwidth everywhere
• Instantaneous propagation of service-affecting errors, software problems, mistakes and other snafus
• Catastrophic and instantaneous failure mode at a terminal level and no one can work until the problem is fixed. This is leading many people to opt for a 'thinish' client model with a limited offline capability such as Google Desktop.

Finally, remember that just like outsourcing this is ultimately a zero-sum gain. The first mover gain is huge but as others join you, your cost savings will be matched by your competitors. But if you don't do it, you may find yourself at the back of the pack.