Bimodal Computing is becoming the wrong answer to the right questions


For most businesses these days that old notion of bimodal computing – the two states of legacy applications and cloud services being treated as separate entities – is past its sell-by date. Most businesses are already in a world of infinitely variable modality.

Composite image of maze question mark © vectorfusionart - Fotolia.comIs the notion of bimodal computing now irrelevant, or was it indeed ever relevant to the user community?

That was the subject of a panel session hosted recently by Rackspace which, if nothing else, demonstrated that the IT industry and its surrounding services continue to struggle to define its problems, let alone find solutions to them.

Indeed, in the case of bimodal computing, it became possible to see that the definition of a problem is being left behind by the user community’s ability to move beyond the problem it set out to define.

Bimodal is, in many ways, even more relevant to the use and development of IT than most of its proponents might think possible,. Especially those that consider it the definition of a simple model of two types of computing – traditional legacy systems and the new, potentially scary, world of cloud service delivery.

In practice, bimodal is starting to stretch in its meaning. For example, is it quite common to her it used to describe different models of applications development, such as traditional waterfall and Agile. The trouble is that its definition, created by Gartner to mean something pretty specific, has now become diffused and irrelevant in the number of meanings attached to it.

In practice, what we are now talking about is a world that offers an infinitely variable modality that can accommodate whatever is appropriate to achieve a business goal. And that does include everything from old PDP11 and VAX applications which are business imperative but where the source code is long lost, through to bleeding edge processes that are only needed for a week and might not even work first time.

The panel, which was chaired by  UK analyst Jon Collins, consisted of Simon Crawley-Trice, Rackspace’s Director of Global Solutions and Services, Spencer Hudson, Global Technical Solutions Manager for GHD, Bernard Drost, Vice President and EMEA MD for Cloud Technology Partners, and Paul Pogonoski, Director of Cloud Foundation Services with Capgemini.

The key factor here, of course, is how business demand is changing the pressures on IT, for this determines just how many `modes’ it needs to work with. As Hudson pointed out, this is still an area where business managers can demonstrate a lack of understanding and experience:

You would have hoped that by now perhaps the management teams we come in contact with would have more experience on how to leverage IT. One manager might have more experience from using a mobile phone, so he’s IT literate, but there is a dearth of knowledge from an IT standpoint and from a user standpoint. That is the dichotomy, that lack of skills and understanding. It’s a hard sell in the business.

The answer, he suggested, is for IT to use some of management’s favourite IT tools against it. For example, analytics is the cornerstone of what can change the mode for managements, as it is usually difficult for people to argue the case against facts and information. If the latest application development results in producing ten of something, and it was five, managers can see the benefits of what is going on.

Over at Rackspace, providing the infrastructure and environments required for such users now means all areas of its operation require the skills to be multi-modal in order to meet users’ differing needs, as Crawley-Trice observed:

My team are the architects designing it. Over time we had to transform to get the right teams in place. On Amazon Web Services and Microsoft Azure, they have separate teams supporting those different business units. Also, customers demand different platforms from us, for they may run their own datacentre today, but want some agile platforms with Rackspace. We have customers running both. It depends where they are from a business perspective.

A similar view came from Bernard Drost, who noted that big financial players now have to learn to be multi-modal and adapt fast to changing circumstances, while at the same time keeping the established servers running:

In financial services, you have all these new up and coming payment platforms. Before you know it, they’ve taken a fifth of the market. You have to pump out features really fast, but keep the main infrastructure going to pay the bills.

From Pogonovski’s perspective, IT has finally felt the impact of the democratisation of enterprise IT. It took a foothold with mobile applications, where companies could start to develop applications without reference to the developer community. In addition, the rise of AWS has meant that, suddenly, anyone can do it themselves or engage someone to give them computing power, even for established backend systems.

So these days, any model of modality is only transitional, as Hudson observed.

Lots of people are doing this but they don’t know it. But the opportunity I think people are missing out on is looking at a process. The destination of a new website is immaterial. There will always be a new website. What people aren’t doing enough of is enjoying the journey. I genuinely mean it. We’re not optimising it enough. The caveat is, with big data and analytics, people are getting really interested in the fact you can measure this stuff. That’s the one ray of sunlight. Even those who are not seeing it, it will come and bite them in the ass.

Even shadow is just a mode

According to Drost one of the key factors is to make sure there is full transparency on what IT is being used and for what purpose. For example, he sees many organisations using – or allowing the use of- shadow IT, with five teams running different things, and all doing it differently. They are, therefore, paying five times for the same thing. The key, therefore, is to keep the base of the infrastructure stable and understood.

This also helps businesses with the applications development process because it allows them the chance to fail fast. This, of course, is now one of the core philosophies of application development and, through that, of building new business processes.

To do this effectively what is important now is for businesses to be able to appropriate their IT infrastructure and resources and to be able to provide whatever is required to deliver whatever operational result and value is considered appropriate to the needs of the business at that time. This is where the increasing use of cloud services can play an important role, of course, for it provides both flexible resource availability and the agility to exploit it rapidly. That is the key to having an infrastructure with infinitely variable modality.

Another aspect of this is the changing of the ground rules as they apply to application development itself. Hudson observed that the cloud had created an environment where individuals need not know about the technology per se, but need to understand how to get the value they seek out of it. This marks a potentially important change, for it shows users looking at the technology much in the way we tend to view cars – a means to an end of getting from A-to-B by using a defined process known as the Highway Code.

The coming of the cloudy cohort

There was an acknowledgement that there is now a change coming about in the general CIO community, where a new cohort is coming through with experience that is based on using mobile, social media and cloud technologies. As these people reach the level of CIO, the fact that they have never had the responsibility of ‘owning’ an infrastructure is starting to have an impact.

This is now an assumption which should be questioned, perhaps even for well-entrenched legacy applications. Such thinking automatically opens up the range of modalities available to a business, as Pogonovski observed:

Businesses are fundamentally changing, and the cycles are changing. In finance, a competitor might bring in a new financial product. Those established businesses then have to compete with that. If you have an established development timeframe, you’re probably six months or a year away from even starting to compete. These days, companies must engage with customers the way the customer wants to be engaged with.

One of the important first steps in achieving multiple levels of modality is creating an environment where starting any development – with all its attendant risks of failure – is seen as `a good thing’. It is certainly a model that Hudson is building:

This is about having enough rope in the organisation to try and risk new things. You need senior support to do that. You get there by running enough projects successfully to garner trust. There’s nothing more to it. You get enough trust that they allow you to go off-piste a bit. Then you go off-piste more and more, but you always deliver. Then it’s not me who has succeeded, but my manager.

The criteria for success, he suggested, is to start incredibly small, perhaps with just something like a Word template, or restructuring your meeting:

Try something. The status quo is such a hard thing to push against. You have to step out of that and try something.

My take

This Rackspace exercise demonstrated an important factor in any discussion about IT these days. There is now danger in accepting definitions of what is happening in the IT world as the gospel truth. Bimodal is a good example, and it has stuck to the point where users come accept that there are only two ways of ‘being’.

In practice, the combination of cloud, mobile, hyper-converged, DevOps and the rest means that there are now as many modal options as there are brains to think of them. The one important mode for all businesses to adopt is based around the impact on business. Shortening the time to cash, for example, is a much better objective than trying to decide whether technology mode 1 or mode 2 is the better option.

Image credit - Composite image of maze question mark © vectorfusionart -