Dr. Hossein Eslambolchi
April 2012
Things were reasonably stable in the good old days. Following some of the biggest challenges facing technology companies were competing against a few, relatively slow moving industries. At the time, the advances in these industries were rapid.
The telecommunications long distance business is a good example of a slow moving, yet extremely effective business. Every long distance service provider had to interconnect its network with one or more access providers to be able to access homes and businesses. Prices were fairly stable – there was a planned market share loss in the transition from the old monopoly to these more competitive new industries. These were the best of the old days. Service providers liked the old world pace of change — progress was very methodical, and there were only a few companies that could compete in the industry. Long delivery intervals — as long as 18 months cycle time — were the norm. Clearly, this is not acceptable in the 21st century, with the advent of Internet based systems. Time to market must be much more aggressive in order to compete. Before the Internet, service providers controlled their services from end-to-end, with no major competition or applications (think over-the-top players, like Netflix and iTunes). This is no longer the case.
During the early part of the 1980s, we had a high-quality, reliable closed loop system. It was stable and provided excellent quality of service. Even though network operation and maintenance were labor intensive, no one paid much attention to automation. In fact, it was not until the early 2000s that I began using what I call the “Concept of One” and the “Concept of Zero” as a foundation for automating every human-to-human and human-to-machine interface had a vision of full automation and an autonomic network, customized by customers worldwide.
Since 1984, things have changed significantly. Networks that are under construction have few, if any, of the features that were built-in in the 1970s and 1980s. With the Internet and all the applications that can run on it, it’s an open world. The Internet has created a situation in which the telecommunications industry is on a whole different course. Innovations come from many quarters and network services are most often developed by others players in the network ecosystem. Networks are now simply enablers that make services happen — not the key and core of the service provider’s business. Customer requirements are at the center of this universe and networks are enabling these requirements to become a reality in life and business.
The question is what is this new world going to look like for the industry? I believe that the most important thing we can do for our customers is to provide high-speed, high-performance access to this open, innovation-driven environment. Mobility can be added as another major end-point to deliver services.
This is why companies are investing both in Internet and mobility technology like never before; their aim is to deliver state-of-the-art applications in support of customer needs. But these investments also create massive complexities for connecting new world technologies to old world technologies.
For example: Consider launching a call from an IP starting point to a circuit-switched end point. From the system and network perspective, the resulting complexity becomes greater than most people imagine. That is why it’s difficult to retire old systems, networks and applications and move to new state-of-the-art models.
This problem can be solved. At AT&T I used the “concept of one” and the “concept of zero” — automating everything with a service-oriented architecture and then moving to quickly eliminate old systems and applications. This was especially challenging in the transformation of systems that rely on main frame computers, like billing. One step that many companies fail to execute flawlessly is placing someone in product management role to simplify the product. Simplification speeds up the development process and drives lower cycle time and increased revenue. I propose that each company needs to have what I call a simplification officer in charge of simplifying products and processes before automation starts in earnest.
In 1995, I predicted this transition when I declared that IP will eat everything. One can get much better QoS with MPLS and IPV6 — especially if the intent is to stream sports or other events in real time. These services will dominate Internet traffic over the next few years.
I also believe that the key to the future will be the ability to leverage the innovation that is coming into each home and business over the next decade. We need to eliminate the notion of “not invented here”, and augment internal innovations with outside innovations to deliver world class services in half as much time. We have the opportunity to re-engineer all of our support systems following TMF standards — so that they become effective and highly efficient, leveraging the manually intensive support systems that we still use today in the circuit-switched world.
You may have heard of Clayton Christianson, a Harvard Business School Professor that has written a book called the “Innovator’s Dilemma”, in which he makes a sound but sobering observation. He points out how difficult it is for very large and successful companies, such as GE, IBM, AT&T and Verizon, to properly leverage new technologies – particularly what he call “disruptive” technology. For Professor Christianson, disruptive technologies not only make quantitative changes in performance or price/performance, but it also qualitative changes. A disruptive technology is an order of magnitude faster, cheaper, and so on – it changes the entire industry. A prime example is the technology Steve Jobs introduced through Apple products. Over the past decade, Apple technology changed the world — and it followed Professor Christianson’s definition of “disruptive” to a tee. Apple technology has already changed every aspect of our industry, down to the consumer level.
IP, Mobility, and cloud computing are probably the most “disruptive” examples in today’s telecommunications industry. They’ve changed everything and will continue to do so for many decades to come. My prediction that IP will eat everything is finally becoming a reality.
What is different in the new, IP-dominated world? The competitors of the circuit-switched world are now working together. IP has absolutely zero boundaries and it renders networks truly open. Ownership by any company or any government on the face of planet is no longer feasible. It may be the only technology in the past 200 years with such attributes.
This is yet another major impact that the Internet is has had on the businesses in the 21st century. No one can debate this, given the massive infrastructure and important businesses that have been built in less than two decades.
Given this general context, there is a variety of other technologies which are changing different industries worldwide. Optics technologies are performing more and more processing at the optic layer. Once networks are re-engineered worldwide, data centers will give us the ability to take advantage of new optics technology.
For example: One can use cloud computing, combined with the “concept of one and zero” to automate every human-to-human and human-to-machine interaction. Networks will become autonomic and computers will be in charge. With the use of data at the packet layer, along with business intelligence and learning methodologies at the network layer, it is possible to create fully autonomic networks.
Just as I once predicted that IP will eat everything, today I predict that wireless will rule the world. Today’s customers have multiple wireless devices and smart phones; connectivity is going through the roof, with both end-points able to connect and generate traffic. This is a great opportunity to monetize the network. The latest wireless technology, like LTE+, is all IP-based. We end up with IP-to-IP experiences through applications such as social networking, Facebook, Twitter and many others. [WHAT] will find its way directly into enterprise markets with massive monetization on the horizon. People think this industry is mature — they believe that wireless is more alive than ever in telecommunications.
I would beg to differ, as new monetization opportunities have yet begun. Once we are able to use deep packet inspection (DPI) throughout the network and apply business intelligence to the resulting information, it will be possible to generate new services and revenue for the entire industry. By 2025, we could have well over 250 billion wireless devices connected to the Internet. That will create substantial amount of intelligent information from DPI, but also a world more vulnerable to cyber-attacks. Security and privacy issues are critical areas that need to be investigated.
Finally, a few words about innovation in speech processing – It may not seem at first glance to be a disruptive technology. Considering Moore’s Law, we may have second thoughts. Moore’s Law predicts improvements by a factor of 60 every decade in price or in price over performance. The laws of photonics predicted a growth factor of more than 100 in 10 years in speech technologies such as Automatic Speech Recognition (ASR). In the past 5 years alone, speech technology has made improvements at a 250 fold rate! That’s four times the rate of improvement over photonics – a rate of improvement that allows us to seriously consider replacing manually intensive customer support, provisioning, and reservation processes for airlines with completely automated speech recognition applications.
Siri is a good example of these developments. In the near future, language translations will be performed in real time, using innovations in speech technology. Consider a call from an Italian speaker to a Spanish speaker, with no human translator in between — this will soon be possible through the massive amounts of computing power available.
I don’t mean to minimize the challenges, but the point remains — existing technologies can translate over a one million word vocabulary system with a very high level of accuracy. Computational linguistics, text-to-speech, and a variety of other technologies that can create an automatic speech system are here. They will be increasingly deployed in networks and applications over the next few years. The dream of talking to computers and using them to improve our daily lives will soon be a reality — similar to what we once saw in Star Trek episodes. I believe speech technologies will be truly “disruptive”, as defined by Professor Christianson.
My To Do list is simple, yet profound.
- Deploy innovations needed to access networks to IP, even if it means replacing them with LTE+ that provides over one Gbps of capacity. Access investments are very expensive and difficult to justify, but necessary to support the demand by most customers who have growing expectations for more and more bandwidth. Moving to IP will allow everyone to innovate around it, including all access providers.
- Cable needs to transition to CMTS 4.0 with speeds of near 1 Gbps, which can easily compete with any access provider on the planet. This is a great opportunity to build major revenues for cable companies in small and medium size businesses. Along with enterprise businesses, there may not be a good business case to deploy fiber to every building in the country, due to its expensive capital requirements.
- Create a new service infrastructure like SOA – make everything e-enabled, period — no human intervention in between, like the Amazon model, end-to-end.
- Work much more aggressively with innovators around the globe. I call it “turning the network upside down”. One of the most striking characteristics of the new Internet economy is the rate at which ideas surface, deploy, succeed or fail. We have no choice but to work with other innovators. The best examples I can think of were Steve Jobs and Bill Gates, who each created one of the world’s most popular products. They were successful because they had tens of thousands of innovators trying to make their products deliver great solutions for customers. They also made sure that they delivered the most popular products. Any company that builds on an IP platform will be a strong winner in the 21st century, and that includes my friends at Google, as well.
- Transition to IPV6. Even though we all love the Internet. I predicted the magnitude and growth of IP in the 1980s. However, the Internet was not designed for the needs of 21st century applications, so a lot of effort will be required to transition the Internet to the new world. IPV6 is key to this transition, which is necessary for all systems and applications, both at the access and network layers.
In conclusion, I believe that the new Internet must support Sextuple play: Voice, Data, Wireless, Video, Gaming and sensory capabilities. It has to be optimized for three major parameters of IP: packet loss, jitter and latency.
Once these issues are resolved — without applying a Band-Aid to the problem as we do with current security methods — the vision I’ve painted will finally materialize. Star Trek will become a reality. Although no-one can beam up Scotty yet, there are no laws of physics that are violated in building that concept!
I am, if anything, an optimist. I believe that human teleportation will happen within the next 2-3 decades. People reading this blog may have a good laugh when it does.
