Tuesday, 16 November 2010

Bandwidth crunch: myth or real threat? 3


Another post about the so-called "bandwidth crunch": an interview with MIX, the internet exchange point of Milan.

1) a ABI research in august 2007 (quoted by many sites afterwards) said: “The increasing bandwidth demands on cable operators will soon reach crisis stage, yet this is a ‘dirty little industry secret’ that no one talks about.”
What does MIX think about it? Is a "bandwidth crunch" a real threat?

Well, this is not a question having a simple answer beeing strictly related to the complexity of the Network and also the the ABI research statement would have to be read in a more articulated way.
Let's have a look at the macro Internet infrastructure: it can be roughly modeled into two main parts; from one side we have the access networks (the network components facing the end user, such as the DSL access networks, the wireless access stage for mobile users, the wi-fi areas); at the other side we have the long-haul transport networks and the network to network interconnections infrastructures where the IXPs like MIX play a major role.These two environments may belong to different operators or large operators may provide them both, according to their size and coverage.

Despite MIX and other IXPs experience a stable and strong traffic growth year after year , we do not foresee any significant bandwidth crunch problem here. Fiber transmission technology is progressing at a pace where long haul connections can accommodate even the most optimistic growth, where high speed Ethernet environments like IXP - even if 100 Gbps will be delayed till 2011 - will not significantly suffer from it while the traffic we expect will still grow significantly.

If exists a possible threat it could be at the edge of operators' networks, on the access part. End users are eager for high bandwidth while nobody is willing to pay more for it.
In large and industrialized country FTTH access infrastructure is the most reasonable solution. ISP and network providers may rely on this shared infrastructure focusing on services provided to end users.
This requires availability by network operators to long term investments and - maybe more difficult ! - a long term vision.

2) Worst case scenario: a terrorist wants to hit the MIX: if he succeded, what would happen to connection in Italy?

Internet does not stop to work. It would happen that part of Internet traffic going from one provider to another could flow slower beeing force to be routed through alternative connections not always enough sized. End-user could experience slowing down in accessing web sites but just if they are located outside their providers network. This is due to the fact that any ISP connected to MIX has at least one other connection to the Big Internet independent by MIX switches and MIX DC (paradigma of "Internet diversity").

3) What's the percentage of growth that MIX can handle (in the near future) ?

MIX adopts technologies and equipment able to manage big growth of traffic: at today we are configured for beeing able to manage more than three times current traffic exchanged on our infrastructures.

4) Does MIX think that the present infrastructure (generally, in the world) is suitable for IP-TV?

In the world there are huge differences between countries in terms of access infrastructure and is very difficult to answer to this question without keeping in mind the overall situations.
Countries where the copper lines are relatively short (few km from the homes to the public switch) can provide good IPTV services on DSL lines. Other countries must switch to FTTH or wireless LTE network to offer good IPTV service experience to end users. In general HDTV over IP is possible only through fiber access and this is why the present infrastructure in general is not really ready for that, at least not anywhere.

(Thanks to Valeria Rossi of MIX for the interview)

No comments:

 
- -