Cloud – a time for practicality

Cloud is not new technology, just re-packaged old technology – that’s why it is commodity, because it works and does a good job. That is why the `package’ is the important bit – it is about what constitutes a service, and the way it is both delivered to end users and how they perceive it and consume it.

So the most important development for this year has little or nothing to do with what any IT industry watcher would consider remotely sexy. It is about making the cloud practical.

But then again, the cloud has very little of great technological excitement  – even Apple’s iCloud is in essence a reworking of what has gone before, usually labelled `Google’, but with the increasingly iconic `i’ branding. All the important elements of the cloud depend upon one thing, the use of technology standards, which is the biggest stumbling block to the IT industry’s most favoured mantra – innovation and change (often, it has to be said, primarily for the benefit of the on-going revenue stream).

But it is the standardisation and commoditisation of the technology which has allowed the cloud to emerge and offer users the greatest service – the emergence of an ever greater diversity and granularity of service offerings to suit their business requirements and budgets.

Most user businesses understand where they currently stand in terms of their knowledge and experience in using IT. Most now understand the messages associated with the cloud and how wonderful it can all be. Nearly all of them, however, still have no great idea, and often considerable trepidation, about the process of transition. There are no convenient chrysalis stages into which they can pop, emerging over a weekend break as a fully formed cloud-based business.

So a growing number of vendors are at last cottoning on to the fact that they need to take users through some steps in order to help business users transform themselves. They need to find tools that help them build some of the underlying infrastructure needed without too much thought. And they need to find some service demonstrations, or find areas of commonality with which they are already familiar, to help illuminate that road to understanding where they suddenly utter the fateful words – `ahh, I get it’.

That has to be one of the underlying reasons why IBM is launching a cloud-based Disaster Recovery service at Cloud Computing World Forum forum this week.

DR? isn’t that old stuff? Yes, and it’s not even new as a cloud service. But that does not take away from the fact it is an excellent demonstration of cloud capabilities. DR is traditionally difficult, time-consuming and expensive, which is why many businesses still don’t do it. Putting it out into the cloud as a service makes it much easier to work with and, dammit, attaching it to the `IBM’ brand probably still sounds more reassuring than `Arnold Scroggins Cloud Services’ to the average enterprise.

What is far more important to IBM, of course, is the fact that it has a far richer, deeper and more comprehensive package of services it can offer once a user business has sniffed success with one, albeit silo’d, project.

It is why Microsoft is increasingly attaching its cloud-based CRM and ERP services to its tried tested (and yes, often cursed) Windows user interface. Just about everyone knows how that works, so if it can be used to build and run an ERP system for a reasonably-sized business, the scary old problem of implementing an ERP system (how many noughts after the `1’ would sir like to spend?) just may become another legend of the `old days’.

It even raises the outlandish suggestion that I might understand how to set one up without the aid of complex brain surgery.

It is also why tools that can short circuit setting up cloud services, pushing most of it into the background, are becoming more important. Most enterprises want to end up with a hybrid environment – a mix of public and private cloud services, plus (in most cases) some old, critical applications still run on-premise.

Latency issues are a common cause of the continued need for on-premise services, and they are going to be a major consideration for many users well into the future – unless someone, somewhere does something clever with the laws of physics. There are tools emerging that help reduce latency, such as the new one-box optimisation solution from Blue Coat Systems, but sadly, they can’t make it go away.

But on a wider front it is tools such as those from OnApp – currently targeted at helping the traditional outsource/hosting services providers move into the cloud – which can play an important part here. They are also ideally suited to the need of enterprises – and more specifically the big vendors and service providers that are looked to by enterprises to help and guide them across the divide between the on-premise of today and the cloud of tomorrow. Those enterprises nearly all want to start their cloud transition in a small and controlled manner, building private clouds at first and then, as experience grows, starting to dabble with integrating public services as part of a growing service mix. But they always face the vexed question: `how on earth di I set that up?’ So a packaged toolset could be just what is required.

The switch to a service culture is also changing some fundamentals that are important to building practical clouds. Take security as an example. There is a switch from just defence mechanisms to a more subtle use of complex policy management and monitoring services to identify unauthorised activities. But even that may not be enough.

I did notice that the recent CIA Denial of Service attack prompted at least one company in the monitoring business to suggest it could have spotted the transgression in real time.  My knee jerk response is `good, but not great’.

Stopping such events as they start to happen – traffic management on steroids if you like – is now an important part of every cloud service. And that has to include not just stopping a particular, malicious event but also allowing all the other positive business activities to continue. Something like NetPrecept’s cPEP technology lays claim to just such a capability. It can, for example, filter traffic so that a business can set differing priority access levels to different types of customer or partner – Gold, Silver and Bronze for example. And as part of that capability can also spot, and filter out, a DoS attack while leaving the good stuff to keep flowing. The service level to customers might slow a bit, but it won’t lay down and die.

Not the greatest level of excitement to be had amongst this lot, it is true. But it does represent the real world for business users a great deal more than the latest gizmos from the fevered brains of technologists. And if the cloud is really going to take off, this has to be the year of boring and sensible.