CommLedge

Welcome back, IA. After more than three decades of doubts and long years of purgatory, it seems that IA has got new traction and has become able to finally address some issues more efficiently and less costly than more traditional approaches. Amongst others, the demise of Expert Systems and their failure in addressing more than very simple problems (and their absolute lack of edge against algorithms) has been part of those failures that have cornered IA in general into the space of promises never held.

Beyond the spectacular but not really significant applications like Siri (and its capabilities of Natural Language processing - at cost, considering the vast amount of servers dedicated to it by Apple), new IA-based capabilities have now emerged that make it possible to offer significant new features, eventually differentiating.

We will have to opportunity to discuss this in our blog posts, but it is making sense to mention at least one example: it becomes possible to consider that language translation is a realistic application domain in the short to medium term. We are not talking there about Google Translate or that class of applications, but more complex systems based - in particular - on relatively specialized semantic networks. In their domain of expertise, these systems are able to produce significant results and will soon offer a level-playing field where translation from currently 'exotic' languages to more main stream ones will create new opportunities. Think about the possibility to accurately translate to and from Chinese and how it will change the approach of smaller players all over the planet to the wealth of information that is today encapsulated beyond obscure ideograms!

The emergence of new business models in ICT has started a while ago. One of the very visible ones was established by Google in a pretty successful attempt to rack the money from on-line advertisement. At that time, it was relatively easy to garner very spectacular results in a fully new market. To be fair, Google and its followers have not just racked the money, but they have built a very efficient machine that has also expanded the market to heights unforeseen at the beginning. To a large extent, they have been the true heroes of the adventure of creating value from the internet and the web, much more than the Amazons who have largely transposed an established model on-line.

The difficulty has come with the two-digit expectation on the growth of on-line revenues. Another model (that is also an extension of the advertisement one) has emerged around the appropriation of user data value by new players against more or less useful services. The obvious example is Facebook (and the many likes) and the provision of its 'narcissism engine' to over a billion of human being across the whole planet (narcissism being apparently a global value).

But it turns out that many signs of the crisis of this model are appearing, one amongst others being that Facebook starts to loose customers from its early adopters base (especially in North America) in significant numbers. The current business models are rather unstable, far away from the stability (and zero growth now, at best) provided historically by the subscription model of the Telecom Service providers. They may also have become unable to provide continuing fast growth as expected. To overcome these limitations, new models are under test, lots of changes ahead.

To no one surprise, cloud is in this list!

Actually, there are many ways to look at Cloud Computing as 'the' major change that is going to happen to the ICT industry and, beyond this, to most everybody on the planet who happens to be connected to a computer-like device (i.e. considering the fantastic deployment of smart phones, a significant portion of mankind).

At this point in time, the 'cloud revolution' has just began and the market addressed today by cloud applications is still very small. One major part of potential cloud customers - Enterprises - has not yet really fully balanced the benefits they may get from transferring their applications from inside to outside in the cloud. A reason for this is that the benefits are not that obviously visible yet, except for a relatively specialized and well touted subset of applications (especially around IaaS). Some difficult issues remain to be addressed, like Service Level Agreements (SLAs), before an open choice of providers with clearly understandable policies emerges.

But cloud is not just for the Enterprise and has profound impact on people too. Another important aspect of cloud is the change in the property regime for the customer. ICT has developed around the purchase of products and is turning to renting them (and not granting anymore any property right to the user). Consequences are becoming visible and will be very interesting to follow. We will!

Though not yet clearly, this is where the most fierce battle is raging. In the last 20 years, the ICT sector has introduced new ways to innovate in almost every aspects: R&D ecosystems, technology acquisition and transfer, business models, design techniques, development techniques (for software and hardware), you name it ... No such example of rapid transformation of an industry since the advent of electricity at the end of the 19th century. The pace of change was hard to follow and that was only for the 'happy few'.

For historical and cultural reasons, the heart of innovation has been beating in Silicon Valley (and in the USA at large) and its core engine has been defined in the Valley terms, with a well established Public Relation engine (technical sites and blogs, acclaimed heroes and geeks). Now that the barriers to entry are lowering (every garage in the planet can become a software factory) and that the 'cultural divide' is reducing (and one day the impact of the 'Shanghai Ranking' in this process will deserve careful examination), other hearts will emerge. This is just a question of time. Looking at the details, we can already see new examples of spots in the planet that are unexpectedly at the forefront of innovation in some specific domains.

Not a day without a flurry of papers on the latest (more and more boring) developments of the epic battle in the Court. The 'patent war' between Apple and Samsung (and some other contenders which have already tried to leave the battlefield with enough feathers left) has been presented as an example of the importance of 'ego' (in this case, Steve Jobs) in decision making.

Actually, this is wrong by many aspects. First, one has to remember the impact of Windows as a copy-cat on Apple in the 80's to give credit of some rationality to Steve Jobs. More importantly, this new patent battle has started essentially because of the need for new entrants (Apple, Google, Microsoft to some extent) to establish a solid position in an Intellectual Property field (i.e. telecommunications) that was governed by bilateral agreements on licensing terms that had taken some time to emerge and granted a relatively good level of protection to the actors (large and small).

The new entrants needed first to establish themselves swiftly in the deterrence dance (don't bite me, I shall bite you more) and in the definition of the Intellectual Property regime. The fact that the current regime was FRAND (Fair, Reasonable and Non Discriminatory) rather than RF (Royalty Free) was a secondary issue.

At this point of war, it will be interesting to see how long the battle will last (we predict: not so long, heroes will soon get tired) and whether or not litigation will continue to be a commercial weapon (we predict: the fight on innovation will prevail).

The promise of Internet of Things (IoT) has been made some time ago and has taken time to become reality. We believe that IoT is emerging as a technology that can be deployed on significant applications. Of course, many challenges remain, in particular regarding the possibility to cost-effectively connect myriads of objects interacting through a plethora of protocols. What is changing is the advent of cheaper 'objects' (chipsets, devices, etc.) and the emergence of platforms (SDKs, protocol converters, etc.) that allow to focus on the applications, not on the glue between objects.

IoT can be, in principal, a domain neutral technology in the sense that it is possible to use the same underlying Machine-to-Machine (M2M) service platform for different industries (this is, for example, the approach taken by the oneM2M standardization consortium). But, like many other ICT technologies that can apply to other industrial sector, IoT may be seen by the incumbents as a 'troyan horse' for new entrants, thus requiring that a specific flavour be developed for a specific industry. This may be a negative factor (in particular from a cost point of view) to the deployment of new applications involving new objects. Following the progress in this field will also give indications as to the maturity of the IoT domain.

And the IoT reality will be real when issues like privacy (or security) will be handled as the cornerstone for success. A very important evolution we will follow.

Regulation is for suckers. Some say. They are wrong.

The general feeeling about regulation in the last three decades was that of a ‘necessary evil’ that should be cornered efficiently and not given a chance to structure an industry against the ‘laisser-faire’ approach that was deemed to be the only one to follow at the best interest of its main actors. There are some lobbies behind this demise of regulation as a means to support the business, but there is also a kind of implicit - sometimes naive - vision of a self-organizing, benevolent community that may solve the problems without government implication. As an exemple, to some extent, the Open Source proponents are often sharing this belief.

But some zones of incumfort are becoming increasingly visible and will likely require a greater involvement of the regulators, nationally, regionally and world-wide. To name a few : net neutrality, internet governance, privacy.

Epic political battles ahead. And a potential game-changer not just to the benefit of people (this will remain to be seen - and our blog will surely address this at length) but probably also for the ICT industry itself.

Blending the most advanced ICT technologies within an industrial sector with low-intensity ICT will give birth to new applications and systems that may change our lives very profoundly. Smart X: here is the new complexity zone. Smart Grids (now), Smart Health or Smart Transportation (soon), Smart Cities (in the longer term) are the best examples. This next generation of global systems is ‘about’ to emerge. What is meant by ‘about’ may vary from one domain to another, with the complexity of the problems addressed and the possibility to deploy applications incrementaly.

This is a dual challenge. For the industrial sectors concerned that experience lower barriers to entry and the advent of new competitors bringing new business models. For the ICT industry whose technologies may be challenged by much more stringent requirements (reliability, efficiency, security, privacy, etc.). For both, with the need to bridge the cultural divide and invent new environments, ecosystems.

In short: an opportunity for ICT, but not a triumphal route either.

Yet another seemingly bizarre item in the list. And a strange way to end it, apparently.

This being said, unless you are a arch-dominant player in a world of silos (which is rarely the case for anyone now, even for the big ones like Microsoft, Apple, Samsung, etc.), most everything will end up in standards before you can do most of what you want. Just to take a simplistic example: it takes over 200 standards to make an iPhone and start to compete on specific, differentiating features with competitors whose products also include the same 200 standards. So what ? Actually, here again the situation seems to be changing, even if less rapidly than in ICT, consensus building still being the norm (!) and taking time.

In the last ten years, ICT standardization has lost ground for two main - and related - reasons. On the one hand, the software product life-cycles have been drastically reduced, not leaving time to go through long processes for creating consensus-based standards: it was the time of mushrooming for lean and mean, narrow-mission-oriented standards organizations.One the other hand, the creation of eco-systems around large companies (and their associated environments: SDKs as well as Intellectual Property regimes, etc.) have favoured ‘semi-proprietary’ solutions over generic ones.

This trend may stop, first because of the cost of large and dispersed standards efforts for companies, but more importantly because the need for large and complex systems (as pointed above as Smart X) drive strong requirements on ICT technologies and architectures and thus make a better case for standardized solutions. To be followed (and CommLedge is involved there).