What next as cloud grows and on-premise starts to fade?

A picture of the change in datacentre coverage has emerged from the Uptime Institute, where its latest datacentre census indicated that a significant shift towards third party service providers of all types is now underway.

At the same time, there are also growing signs that the third party service provision community is trying to work out what to do about the big gorilla in that marketplace – Amazon.

The Institute’s findings, based on a survey of 1,000 datacentres around the world, suggest that over 60 percent of the third party service providers, which covers those offering cloud services, hosting, and co-location facilities, had budget increases last year of 10 percent or more. In other words, the majority were upping there spending to accommodate new business.

Enterprise datacentres, on the other hand, did not fare so well. Only 25 percent had a budget increase of 10 percent or more.

Cutting the cake another way, it also found that just over 20 percent of enterprise datacentres were suffering budget cuts, while only five percent of third party providers faced any cuts.

That all points to an increasing amount of the overall workload being outsourced, at the very least. This suggests that even those businesses which are outsourcing to third party server farms are also now putting themselves into a position to at least start moving towards establishing private cloud environments.

This trend does open up an opportunity for the third party service providers to start attracting those users to think in terms of more than just hosting resources `somewhere else’. It is a step that is being taken by some of the established systems and software players such as IBM, Fujitsu, Oracle, Tibco and the like. They have cottoned on to the fact that providing more comprehensive services to users – services that mean users no longer have to think about specifying, purchasing and implementing those services in a cloud environment – can help to make their business more `sticky’ to their customer base.

The primary attraction for users, however, is for now still price. The `services from…….’ price tag marketing approach is still one that sells well in the third party service market, with Amazon’s AWS being the big name in that game.

The company cut its prices earlier this year, and now probably has a strong enough hold on the hearts and minds of its customers to be able to beat off most competition looking to play by Amazon’s rules. It will, for example, be interesting to see what happens with US IaaS provider, ProfitBricks, which is now setting out to do just that.

It is attempting to pre-empt the next Amazon price drop by dropping its own per hour prices on resources by upto 45 percent. This brings the price of renting a one-CPU storage server with 3.75Gbytes of memory and 410Gbytes of  storage down to $0.1044 per hour, versus $0.12 for an equivalent system on Amazon.

That is likely to be a hard row to sow successfully, even though its target market is only one part of Amazon’s total market view, the developer community and others needing resources for short term projects that don’t require reserved server instances or long term contracts.

Attempting to beat Amazon on price is likely to be a fool’s errand, and one that even the company is aware of, as it starts adding services to its base offering of rentable resources.

The one problem here is then trying to identify what additional services and technologies should be bolted onto the resource services in order to attract customers and make services sticky. This is the issue now facing Rackspace Hosting.

As the Uptime Institute numbers suggest, Rackspace Hosting’s cloud business is growing, and growing much faster than its overall business. The trouble is the company's overall growth in sales weakening and its profits are shrinking. That tends to be what happens if a company takes on the big player in a market on the basis of price alone. And it is an approach which will not really help the company too much. It is already seen as Number two in the market to Amazon, but is a long way behind it.

The company knows it needs more, which is one reason it is the lead house behind CloudStack, the open source infrastructure and platform technology which has the potential to be one of a few leading platforms on which users can easily build the service architectures they require.

That is exactly the type of development from which sticky customers are made. Get that platform right and the services which can be built upon it become what the customers are really buying. Those services are geared to the sellers’ perceived expertise in the customers’ business needs – and the technology can become `invisible’.

The one issue the CloudStack movement has here is that history shows that such accepted standard solutions come from user acceptance of a proprietary product more often than from a specifically tailored, community developed, standards-based design process.    

But opportunities are around, and more to the point, the company is aware that it needs to come up with new ideas.

For example, while its dedicated hosting business is slowing, it is still very much there: and those customers will be the ones that are likely to spend well when they make the move towards cloud services. The company is now building tools that will help dedicated hosting environments move into cloud infrastructures. It will also then be in a good position to integrate those cloud with any parts of a company’s service requirement where running applications on bare metal services still makes sense.

The company’s resident cloud evangelist, Robert Scoble, has also called on Rackspace to launch itself at new areas for development, such as meeting the needs of business to personalise services and build new collaborative enterprises, rather trying to gain technological parity with Amazon.

Scoble is concerned about an innovation gap with Amazon and there is some point to such a concern. On the other aside of that coin, however are two issues. First of all, in a still juvenile business like cloud, innovation gaps are going to exist all over the place – and most of them will prove to be transitory, and pretty irrelevant over the long haul. That leads straight to the second issue – what constitutes the right type of innovation.

As well as pointing to collaborative enterprises, Scoble suggested wearable computing devices as key development. Up to a point that is likely to be so, but is that an appropriate area for development by Rackspace? Somehow it seems unlikely, whereas developing the APIs that allow any wearable device to collaborate with any appropriate application or service, is where its development skills would be far better directed.

By the same token, building the tools that allow collaboration of any service with any service, regardless of CloudStack, would be a capability which would make Rackspace, a very sticky company for many potential users out there.