This post is about cloud computing (referred to simply as cloud in this post), focusing on storage and content.
The entire ITC industry talks cloud, so let me try to give my perspective. For customers the debate, at this stage, can operate safely on a hypothetical level exploring what this buzzword might entail. For people in the industry, most of my conversations are either about what their respective corporate message dictates or what they as individuals speculate could change the landscape. Martin Hingley, on January 11th, stated in his blog that 2010 will become a profitable delivery mechanism. And Martin is specifically addressing what to expect of the growth in 2010.
Cloud implies a mixture of technology, infrastructure, cost models and changes to decision making. Some compare it to the truly revolutionising impact of the internet. I believe cloud is a force for good, but it will be slower and less disruptive in the world of enterprise IT. However, this will not diminish the impact of cloud since it will be widely adopted as the way to build future infrastructures.
Cloud considerations:
The issues to explore are:
Infrastructure: Cloud means having an IT infrastructure based on (1) Internet (2) virtualisation and (3) being highly scalable. The infrastructure can be entirely outsourced, and in the extreme be run by someone who never accounts for how the cloud is done. But those 3 elements will all be there in my opinion. The infrastructure will increasingly be based on standard building blocks, highly networked and there is still plenty of scope for inserting unique IP by the vendors.
Outsourcing: Cloud may be outsourced, but does not need to be. It requires a cloud infrastructure somewhere, which could entirely be hosted by internal IT. Outsourcing is not a prerequisite but could well make a lot of sense, when someone is better and cheaper at running the cloud.
Cost: Cost is a key ingredient even though a cloud infrastructure could purely be justified on a better and more reliable quality of service (QOS). However cost is always important and so is focusing on holding any supplier accountable.
Change: The changes necessary to enable a cloud infrastructure are considerable. Therefore, understanding from the outset that the scope of change implies risk and controlling change successfully is critical.
Control, security and performance: The promise of better & cheaper outsourced IT provision should only be evaluated once the issues of control and security are understood. Some IT applications are so critical that IT departments would never let go of total control. Often this is purely down to the question of where data is stored. Security is a similar issue, which can be solved, but needs to be fully explored before an application is released onto the cloud. Performance is also a unique requirement that will dictate whether an application is a candidate for cloud, or not as the case may be. All these are not specifically cloud inventions, but cloud makes them critical decision making points.
Storage and content
So what about storage and content in the world of cloud computing? As stated, first decide whether total control is required, which then opens the possibility for a private cloud solution. Where control and security can be delegated, then this is where plenty of recent cloud innovation has been developed. The leading visions for storage proposes that the cloud infrastructure can reside anywhere, co-exist across several tiers, and that applications/users connect to gain the required access to the content.
So, the benefits of a cloud environment can span from just exploiting the foundation of a cloud infrastructure all the way up to a solution that is completely outsourced. Not unique to storage but distancing yourself from the content is a difficult decision to many IT people.
One obvious candidate for cloud has been the area of backup and business continuity. In fact the ability to provide these at separate sites have been a driver for cloud-like solutions before cloud was defined. It may therefore be paradoxical that some IT departments would never consider outsourcing their data, but happily place unencrypted backup tapes with a 3rd party. In this case many IT professionals ought to consider a properly secure cloud based model for BC backups.
Cloud growth
Consumers love their free cloud based applications and storage. This is expected to grow further with alternative pricing models in addition to free delivery. Medium sized business and IT are also candidates for cloud based solutions, where in-house IT is expensive compared to hosted solutions. Enterprise IT will increasingly experiment with tiers of cloud computing, but I do not expect cloud to take off quickly. Yes, some will be early adopters and embrace the infrastructure since cloud is attractive but most have legacy environments in place. Applications that turn out to be good candidates for cloud are those where the whole task can be outsourced.
What to expect from vendors?
Many storage vendors are happy to define themselves as the providers of the building blocks for the cloud. Ie. other players typically in hosting or services put complete cloud solutions together. Both approaches will mean increased pressure on buyers with seemingly conflicting propositions of what cloud means. It is clear that cloud has taken off as a marketing concept, jointly we will have to make sense of it. This will consume time and effort on the part of customers exploring what cloud infrastructures can potentially offer.
#1 by cegge on January 18, 2010 - 18:35
Let us consider the challenges of the many over the few although having the benefit of the experience of the big cloud providers might be good. Seriously, and conversely the task of selling cloud to the many is considerable. Vendors need to understand what it takes to convert cloud skeptics into early adopters.
#2 by Martin Hingley on January 18, 2010 - 17:41
Claus
I enjoyed reading your article and thanks for referencing my post.
It’s worth thinking about the comparison between:
· The traditional CIO in a large organisation who has to deliver 100 or more application services to a few thousand users, and
· The CTO of Google, Yahoo, Microsoft who has to deliver a single application to tens of millions of users
Arguably the former has the bigger challenge…. not least because of the differences in software licensing, databases and access technology. Funding is typically harder to get and it’s always harder to change on the fly than to build from scratch.
Data centre consolidation has been a key subject for 10 years at least. Clearly there are now a number of new subjects to address if organisations are going to succeed in turning ‘corporate’ into ‘private Cloud’ computing.
Best Wishes
Martin
#3 by cegge on January 15, 2010 - 09:46
Good points! More reasons why cloud looks like a straight forward strategic decision, but someone needs to pay attention to the details of successful migration. And what looks good on paper may not work in real life. Monitoring is also a good point, but was skipped in the post to shorten it.
#4 by Martin Harris on January 14, 2010 - 20:21
I think that other factors in the cloud space will come to the forefront over time. In particular application scalability, deployment, automatic scaling and finally but not least monitoring.
Not many people realise that you can’t just deploy an application to a cloud and expect it to run happily. Without care it may run but not scale. Applications backed with traditional databases can be a problem, Threading issues and general bad design can lead to issues and outages.
http://www.javaworld.com/javaworld/jw-03-2009/jw-03-multicore-and-cloud-ready-1.html
http://searchcloudcomputing.techtarget.com/news/article/0,289142,sid201_gci1355058,00.html
Deployment is becoming less complex with people like SpringSource making it very easy: http://www.springsource.com/newsevents/springsource-launches-enterprise-java-c
Scaling and generic algorithms that can be tuned to tie in with a budget and profitability. i.e, please don’t scale beyond some kind of profitability calculation.
Monitoring in particular is going to be a challenge, I don’t think anyone is doing anything interesting in this space yet.