Innovation never stops…

Attending Johannesberg Summit 2013 has been such a great experience. Here’re some interesting observations.

What is 5G? When is 5G? Most speakers have talked about their views about 5G and it looks almost sure that 5G will be something exciting ( Gerhard P. Fettweis – TU Dresden – The Tactile Internet,  Lauri Oksanen – Nokia Siemens Networks – Your Gigabyte a day,  Chih-Lin I – China Mobile – A peek at 5G, etc.) No delay, cost efficient, numerous connections, high speed, and so on …. A question that remains open: what’s the fundamental enabling technology?  Let’s take a look at the enabling technologies of existing generations and how long it takes for technologies to be accumulated for each generation:

1985, 1G: analog. AMPS etc. (first wireless telephony 1905, 80 years.)

1992, 2G: digital, TDMA/CDMA etc. (from Shannon, 1948, to 1992. 44 years)

2002, 3G: multimedia, WCDMA etc. (De-Rosa-Rogoff defined the direct sequence spread spectrum

method in 1950, 52 years)

2012, 4G: upgraded form of 3G, rich multimedia, MIMO+OFDM, etc.( 1970,  A.R. Kaye and D.A. George, MIMO; 1970: OFDM, Robert W. Chang. 42 years)

While each generation takes around 10 years to be standardized, it takes more than 40 years of fundamental research to accumulate sufficient technologies for each generation. Utilization and commercialization of research results are getting faster and faster because of the ubiquitous availability of information and the exponential growth of information exchange rate.

An immediate example is given by Lauri Oksanen in his talk “Your Gigabyte a day”. The standardization time of different Gs:

LTE: 7 YEARS

wcdma: 11 YERS

gsm: 12 YEARS

Now after standardizing 4G, have we used out all technologies we have obtained so far? If so, will future innovations also grow exponentially? The bottleneck soon may no longer be channels, but human brains. So are existing technologies sufficient enough to support our demand? Or if we have used out all technologies accumulated so far, shall we be a little bit more patient for our brains to generate some new results? So when will be 5G? And who will be the main contributors?

Another interesting trend is the cellular network is getting more and more centralized (Chih-Lin I – China Mobile – A peek at 5G, Pranav Mehta – Intel Corporation – Building the Software Defined Infrastructure, etc). If you question the complexity, watch Pranav Mehta’s talk and you will be relieved. When we apply Moore’s law also in network computing, all complexity issues will be solved, sooner or later. Then there would be many new research opportunities.

All the talks is to be found at the Johannesberg Summit webpage.

This entry was posted in Conferences and tagged . Bookmark the permalink.

5 Responses to Innovation never stops…

  1. peter says:

    “have we used out all technologies we have obtained so far?” <<— Based on the examples you showed (e.g. TDMA/CDMA/OFDM,…), it seems you define technology as a "radio access technique". In my opinion, that is a very narrow way of looking at things. LTE, which uses OFDM, has a lot of new techniques that were not available in previous generations of cellular systems.
    There is no rule that says 5G must use a different access technique than its precedents. The innovation can be at higher layers.

    • guowang says:

      LTE is mainly based on MIMO+OFDM, others are like complimentary techniques to make using them more efficient, e.g. SC-FDMA, HetNet, etc. These techniques are invented in the past years. LTE integrates all of them that’re needed. That’s why the “accumulation” of techniques is used in each G. Each G is a new revolution of the whole eco system. There have always been higher-layer innovations. If you look at the applications in different Gs, they’re dramatically different. Defining the standards is just the first step of one generation.

      • peter says:

        I still do not fully agree with “have we used out all technologies we have obtained so far?”
        The link-level performance (of a single user) is as close to the theoretical limits as it can be, so why move from OFDM? On the other hand, it is the multi-user case that is still far away from the theoretical limits. This can be solved by higher layer techniques like interference management (e.g. interference coordiantion like ICIC, interference cancellation at the receivers, etc.).
        One problem with academia is that if some ideas or concepts are not published in IEEE then these ideas “do not exist yet”. I have so many times reviewed conference/journal submissions that claim a novel concept which in fact has been presented at one of the 3GPP standardization meetings more than a couple of years ago.

        • guowang says:

          MU-MIMO and MU-OFDMA have been very well understood. Only interference-limited case, i.e. a multi-cell perspective, is way from being clear. But the question would be, will it be possible ever to be clear? Yes industry has complete advantage in technology. Most corner stone technologies in each G, including shannon, are from industry (different research labs, etc). This can be easily understood, experience v.s. newbies, collective v.s. individual efforts. On the other hand academia has the luxury/advantage to thoroughly think about some problems. “have we used out all technologies we have obtained so far?” Well this sentence has a question mark. So it’s not a conclusion. Let industry and academia just question themselves ….

  2. guowang says:

    LTE is mainly based on MIMO+OFDM, others are like complimentary techniques to make using them more efficient, e.g. SC-FDMA, HetNet, etc. These techniques are invented in the past years. LTE integrates all of them that’re needed. That’s why the “accumulation” of techniques is used in each G. Each G is a new revolution of the whole eco system. There have always been higher-layer innovations. If you look at the applications in different Gs, they’re dramatically different. Defining the standards is just the first step of one generation.

Leave a Reply to guowang Cancel reply

Your email address will not be published. Required fields are marked *