Cloud Computing is an emerging style of IT delivery in which applications, data, and IT resources are rapidly provisioned and provided as standardised offerings to users over the web in a flexible pricing model.
At a recent conference where I presented on the current state of Cloud Computing, I was pleasantly surprised at the level of genuine interest given the current low investment situation in which most companies find themselves. I sensed two main drivers here:
The first was that the current hyped IT technologies (SOA and Web 2.0 in particular) had moved from interest, up to over-excitement, through early confusion, up the slope of discovery, and were entering the boredom zone, to paraphrase Gartner’s hype cycle. So, of course, the technorati, with the attention span of a micro-gnat, are keen to look at and play with new toys as they appear. Cloud Computing certainly fits this bill with the subtle promise of a ‘paradigm shift’ (i.e. IT will be asking for lots of money to replace perfectly performing systems with something untried and of unknown cost). It also looks like being this year’s big thing in IT (other than massive redundancies and salary cuts, but I won’t go there this time).
The other driver is that senior executives (who tend to learn most of their IT knowledge on the golf course talking to other similarly IT-illiterate executives – I’ll call them the ignorati) have cottoned on to the cost-saving possibilities of Cloud Computing, both in cheaper processing power, but also (in their simple minds) fewer expensive technorati to run them.
This dual pronged approach is likely to prove irresistible and I expect to see a number of Cloud Computing pilots taking place in the next few months. Note that I refrained from making a ‘pilots in low visibility crashing’ joke then. Anyway, I would advise the usual caution, pinch of salt, bromide in the tea, etc. This is not because Cloud Computing is pants. Indeed it has the potential to be the rare good idea that is also practical, affordable and low risk. The other positive aspect of Cloud Computing is that it is relatively simple to try, provided you try to keep it relatively simple.
For hype watchers, Cloud Computing is shaping up to be one of 2009’s Big Ideas. This is both a good and a bad thing. The good news is that the technologies underlying the deliver of Cloud services is maturing nicely – several, including remote storage (e.g. Mozyhome), collaboration (e.g. Citrix Online), and BPM (IBM’s recently announced Blueworks online BPM workspace) are available and working.
There are two current variations of Cloud Computing currently being tested – public and private.
The public Cloud is seen as an emerging style of IT delivery in which applications, data, and IT resources are rapidly provisioned and provided as standardised offerings to users over the web in a flexible pricing model. Recent announcements from Amazon (EC2), Sun (Sun Cloud) and IBM (BPM BlueWorks) is an example of public Cloud services for software development, hardware and process modelling respectively.
These public clouds mark the next step in the journey to true utility computing and builds on the success of some of the early Software as a Service (SaaS) vendors, such as Salesforce.com and Netsuite. There is now a proliferation of “… as a Service” offerings – Communication, Infrastructure, Platform, etc. – leading to the inevitable EaaS (Everything as Service) acronym being bandied about by some of the more impish technorati.
The private Cloud is a way of managing large numbers of highly virtualised resources such that, from a management perspective, they resemble a single large resource.
This can then be used to deliver services with elastic scaling. All the major hardware vendors (IBM, Dell, HP) are starting to deliver devices and software that helps organisations to test out Cloud Computing in the safety of their own environments while they get to grip with these principles and practices. With the obvious improvement in security with a private cloud, most large organisations are expected to start with their own internal test beds.
From an SOA perspective, the acceptance of Cloud Computing is a vindication of the service based approach and architecture that are the very core of the SOA paradigm. If you want to feel smug, gently point this out to the cloud bandwagon jumpers. In addition, the disciplines of service management, particularly ITIL and the use of a SLA-based Configuration Management Database (CMDB) repositories will also come to the fore to provide the governance for this fluffier new world.
OK, the bad news is that as with all new technologies there are a number of problems or solution areas for which Cloud Computing is not suitable as a replacement for or addition to SOA:
- Workloads which depend on sensitive data normally restricted to the Enterprise, e.g.
- Employee Information – Most companies are not ready to move their LDAP server into a public Cloud because of the sensitivity of the data
- Health Care Records – May not be ready to move until the security of the Cloud provider is well established
- Workloads composed of multiple, co-dependent services
- High throughput online transaction processing
- Workloads requiring a high level of auditability, accountability
- Workloads subject to Sarbanes-Oxley, for example
- Workloads based on 3rd party software which does not have a virtualization or Cloud aware licensing strategy
- Workloads requiring detailed chargeback or utilization measurement as required for capacity planning or departmental level billing
- Workloads requiring customization (e.g. customized SaaS)?
By taking a time-boxed Proof of Concept approach to test its viability and suitability for your organisation will give you an idea of the business benefits, the technical challenges and the risk profile of the service. Even in these financially challenged times, I would recommend that you put your head in the clouds and have a look around. It is better than putting your head in the sand.
