There is an interesting article on the New York Time about Internet Traffic and the relevance (it is just an euphemism) of google in this realm. You can find it here: http://www.nytimes.com/external/readwriteweb/2009/10/13/13readwriteweb-google-accounts-for-6-of-all-internet-traff-90323.html 

What is surprising to me is the absence of any Operator in the list. They once were the “owner” of the core of the Internet (the Tier 1 providers) but probably (as said by B.Norton in many articles. e.g., http://www.pbs.org/cringely/pulpit/media/InternetVideo0.91.pdf) the adoption of a peering strategy adopted by major Web Companies has changed this (have a look at http://drpeering.net/a/Home.html for a deep understanding of what is going on peering). There is the stringent need internally to Operators to focus more on what is going on in the Internet in order to have a role in the entire value chain …

Another point that seems of interest is the Internet transformation. The Internet is becoming Asymmetric, i.e., the huge data centers providing services (google, youtube and the others) are actracting more and more trafic and they are, in some sense, undermining one of the funding principles of the Internet: the end-to-end argument. They are putting together the Server infrastructure with a huge data network. Can we still say the the data center is at the edge of the network? Isn’t it more correct to say that the Data Center is also the network?

Some colleagues at the EU Project NANODATACENTERS (http://www.nanodatacenters.eu/) have already said so and they are working on a highly distributed infrastructure based on P2P technologies (even if the NYT paper claims that P2P trafic is declining).

Peer to peer networks (and in general fast downloading technologies such as youTube and the like) pose another set of paradoxes. The first one can be termed the bandwidth paradox: “the cheaper the content, the greater the bandwidth request”. This takes into account the trend of many users (especially those that are heavy users of peer to peer applications) that tend to subscribe to flat rate offerings of operators in order to have high bandwidth without limitation all day round to use for downloading multimedia content (sometimes infringing copyrights). The P2P or download heavy users are also named “bandwidth hogs” because (even if they count for less than 10% of the total number of users) they are the bigger consumers of bandwidth (they are responsible for much more than 60% of all the data traffic). In other terms it could be said that people tend to invest into bandwidth as a means to collect a lot of content, more that invest directly on the content.  A corollary of this paradox is the storage paradox: it states that “the cheaper the content (e.g., because of file sharing), the greater the mass storage request”. Simply put, this paradox states that users will not invest anymore on the content (that is cheap and can be downloaded in big “quantity” from the network) but on the means to store it. Even considering the more convenient offers from the “legal” market there is a considerable gap between the value proposition of the offerings and the reality. Nowadays there are portable media player devices or hard disks with huge memory capacity (up to 80 Gbyte for mobile devices and in the order of Tbytes for the HDs). If a user would fill the memory up of legally downloaded and paid for content, those devices would store in their massive memories items whose global value (e.g., for portable devices 40 thousands songs or even more at one euro per song) would be much greater then the one needed to buy a luxury car (and usually users do prefer luxury cars!).

These two paradoxes are specializations of a more general content paradox: “the value moves from the content to means to deal with it”.

The previous discussed paradoxes bear to the most important inconsistency (at least from the Operators point of view), the revenue paradox: “the more the network is used, the less it produces revenue”. Currently, users can take advantage of very interesting bundles of service offerings. Many Operators are providing bundles in the shape of “all you can eat”. This means that users pay upfront a fixed amount of money and they can make unlimited national (but for certain plans also international) calls. Also for the data connectivity there is a trend towards flat rates, and in this case too the price of megabit per second is declining very rapidly. Any investment in the network is going to provide little “ROI” (return in investment) because users do not perceive too much value on any network improvement. An example is the continuous increase in the bandwidth provided to users by xDSL. Any further increase in bandwidth can require huge investments in the network in order to augment its capabilities (i.e., moving from xDSL solutions to the fiber to the home one), but for the end users there is no reason for paying more the additional capacity (even if the bandwidth doubles). In other words users are not going to pay proportionally any increase in bandwidth and the price for any megabit transported is going to progressively decrease (while keeping the ARPU stable or even decreasing).

In this context many Operators are called to invest on new (and expensive) access networks for providing broadband connectivity (the replacement of the copper with fiber, for fixed network; and the Long Term Evolution mobile broadband). In this paradoxical situation it is hard to make decisions and planning for the future.

Free Connectivity gives way to another important fact, the pervasive communication paradox: it states that “the more the communication is available and cheap, the more it is pervasive”. This is another way for saying that communication is becoming a commodity. Many users will take the availability of a cheap network for granted. The rich availability of connectivity, the cheapness of networking devices for home networks and the always decreasing price of public network connectivity will allow a great availability of cheap and ubiquitous bandwidth that will be used by pervasive applications that will be “always best connected” both in term of maximum bandwidth and of price.

This is why I don’t think the M2M paradigm pushed by Operators (privileging the SIM based connectivity) will have an easy way. Probably a trade off will be the result: many end points with short range connectivity will send info and data to a few public network connected nodes.

Btw, it will be quite interesting to know what is the possible traffic generated by sensor, actuators and “things” over the public network. Is it something Operators should be ready for or is it negligible? Just think to billions of sensors exchanging data …

The disruption introduced by the “end to end argument” has a heavy impact in strategies and deployments of Operators. But there is even more: “the best network paradox” (http://netparadox.com/), in its concise definition states that “The best network just moves bits and the best network is the hardest one to make money running”. If the network is stupid then it provides simple services that make it a commodity. It is therefore difficult to make money from network services and the only viable possibility is to reduce investments on it providing just the essential functionalities.

Alternative networking capabilities are becoming highly available (from community networks, to municipal networks, from WiFi to Wimax and the like). They could provide connectivity at a cheaper price. Emerging networking techniques (e.g., mesh networking, peer to peer, opportunistic networking and the like) are able to adapt, optimize, and share network resources within communities and they can scale up to many users (in the case of p2p even millions). An example of this adaptation and sharing capabilities is given by FON (What’s FON” available in http://www.fon.com/en/info/whatsFon); the idea is to share the connectivity that a user pays to a Provider also for allowing other peers (members of the FON community) to access freely or for a low fee to the residual bandwidth of the hosting user. Other examples are given by hotels, bars, coffee shops and even municipalities (see the case of Venice plan to offer WiFi connectivity to tourists) that provide connectivity for free in as a competitive means to attract more customers to their premises. In the long run connectivity could be provided for free by many different providers in exchange for the “attention” of the customers. This is the phenomenon of decrease of the “value” of the bandwidth.

The consolidated market of many Telecoms Operators is under turmoil, on one side the Internet technologies are challenging the way Operators create and deploy services, on the other side new Internet business models are eroding the margins of the telecoms market. This situation creates a deep uncertainty and a set of network paradoxes that Operators have to cope with. This and other posts are about a few network paradoxes. 

Many Operators are facing the problem of the obsolescence of their network infrastructure. This problem rises in a new context dominated by the possibility of substituting a consolidated technology (the circuit switching) with a new one completely based on the IP connectivity (packet switching). The so called All-IP Networks (AINs) allow for lower infrastructural and management costs and promise to be the basis for building a seamless service platform able: a) to provide a myriad of new services for the fixed and the mobile environment; and b) to intercept the Internet evolution and its increasing market. However this transition to an All-IP infrastructure has inside the seeds for an important transformation: the consolidated way of providing communications services is going to be fundamentally challenged (e.g., see the post of Roberto Saracco’s about Skype).

One of the foundation of the IP communication (and consequently of the Internet) is the “end to end argument” (see for example http://www.reed.com/dpr/docs/Papers/EndtoEnd.html), which dictates that “mechanisms should not be enforced in the network if they can be deployed at end nodes, and that the core of the network should provide general services, not those tailored to specific applications”. This principle has been well expressed and stressed out by the “stupid network” paradox (http://www.rageboy.com/stupidnet.html). It juxtaposes a slim, simple and cheap IP core network (dominated by abundance of connectivity capacity) with intelligence at the edges to the centralized and expensive infrastructure of the Intelligent Network.

The advantages and the disruption of this move of intelligence are appalling for the Operators: the rich and costly deployed network functions and services are not anymore appreciated by customers that do value services and applications offered by servers at the edge. In this way, the network is progressively emptied of functions and all the value moves to the edge. On the long run, Operators could be forced to play a less significant role, the one of Bit Pipe Carrier and not the desired role of Service Provider.  And this could happen in spite of the significant investments in existing and new service infrastructures (like the IP Multimedia Subsystem, IMS) intended to position Operators as Services and Platform providers also for the Internet market.

This has disruptive impacts on the telecoms industry: the rich market of value added communication services is not anymore under a strong influence of Operators; the existing and the scheduled service infrastructures are rapidly aging or are not fitted for competing with new solutions based on emerging Internet technologies; the telecoms service equipment market is radically changed yielding to a strong reduction of investments (and prices) in the area of service platforms. This course of action has far-reaching effects on the service delivery telecom market as a whole:

• Service Delivery: services cannot be anymore deployed in long cycles (sometimes up to 18 months); they have to be much shorter in order to find a potential market and to compete with the Web Service providers (WebCo) offering;
• Interworking: services offered by Operators have to be readily interoperable and run on different networks. Long standardization periods for new solutions have to be dramatically cut down in order to come out with simpler and quicker deployable solutions;
• Vendor relationships: Vendor lock-in has to be limited; Operators have to be able to really integrate different solutions. In addition Operators should be able to enlarge their ecosystem of vendors sometime favoring the entrance in the telecoms markets of new vendors form emerging countries;
• Equipment costs: the foreseen huge investments in new networks and services require a contraction in costs of equipment. Costs cannot be anymore those of the past but they have to be aligned to Internet prices.