Cloud is clearly the way forward, the ability to scale your infrastructure, your capacity around your usage patterns, budgets and requirements can be a transformational concept. Whether it’s the small business using an Exchange Cloud service for 25 people, or an animation studio using compute on demand services, being able to pay on use, use on your terms and scale up and down with business need enhances flexibility to meet project goals and creates opportunities to test and scale systems or try things out to understand the possibilities for extra functionality, extra services and revenue generation – to online those extra web servers to enable tickets being booked for that new concert and hand them back when bookings are closed for example.
The challenges from an olden days viewpoint:
- We need to understand our business and our application more – what real end user tier is it, none of the usual tier 1 or tier 0 where it can’t go down discussions – not all applications are the same and therefore need specific levels of availability, understand there is a cost to 24/7 never down solutions.
- We need to educate the business teams to be asking the right questions thinking more strategically about their applications and their specific lines of business, with the right set of technologies what could we achieve? How do we translate that to mobile, to make the different apps share data effectively, consolidate applications to reduce cost and complexity.
- We need our applications to be adapted for cloud – more portable, less linked and dependent upon development code versions, JAVA runtime, middleware or specific database and OS patch/feature revisions – more operational within it’s own space and therefore able to ‘request’ and operate with the middleware, OS and database as a client, than embedded into them all. They also need to be more easily deployed – none of that uninstall and re-install, fine if it’s on two servers, but scale that up to 47 and an upgrade becomes painful and a long term activity.
- We need the business requirements defined – availability as discussed above, performance and data processing applications – moving away from a virtual or a physical machine to more defined values, number of users, number of transactions per second or latency.
- We need to understand any application or business data and security requirements so that they can be highlighted and mitigated if possible, or accepted as appropriate.
- We should be examining internal private and public cloud as appropriate, recognizing what our users may already be doing, understanding how we can be agile to users and business needs – make IT become the consultancy, the centre of excellence for our business rather than the supplier of IT.
In essence, IT as a department or a business in it’s own right needs to get out of it’s users way, empower them to make the right decisions, offer support where necessary and operate as both the center of excellence and best practice, as well as the point of escalation.
This represents both a challenge and an opportunity for IT and the professionals operating within it, what is for certain is that times have changed, quoting a good CIO friend of mine “…now you’ve put everything in the cloud and it all works wonderfully, what do we need IT for?”