Wednesday, May 5, 2010

Location Location Location

When you move between different parts of the tech ecosystem you usually end up getting your perspectives moved around, a bit, by each place you have been. In my stint in the Telecom/Networking world I've been through a couple of places, Vendor, Service Provider and now Distributor/System Integrator/Value Added Reseller.

In all of these places you pick up the core values/issues/challenges for the current segment of the ecosystem and if you go through enough places you get decently balanced outlook on the business,
compared to if you stay put in one part of the business. When it comes down to it, its all about understanding you partner/customer/clients business model and empowering them to be successful with your products/services. 

Which brings me to the why I started writing this post, I read this blog post and it brings up some very interesting points and most are probably valid, but the old Vendor Systems Engineer in me cringes at the shallow "bandwidth is not a problem anymore" type casting. 
I keep hearing this from many camps, the bandwidth problem is solved, we will have wireless 100Mbps anyday now, just look at LTE, WiMAX, 4G and 5G and what not.

My problem with that is basically, its not correct, most of these new technologies got two major obstacles, they usually are a shared medium and there infrastructure/core network to support these fantastic speeds are not there and the infrastructure upgrade will be very very expensive and hard to do.
Now one common misconception is that we all get 100Mbps from our local LTE Access Point or some other big number. That only happens if you are the only one connected to the AP and you are basically sitting right next to it. Now if you and a friend are sitting right next to the AP and you both are connected, then you only get 50Mbps each. Which by itself is fine and dandy and it sort of scales down linearly ie add one more user and you divide the bandwidth evenly between the users until you reach the maximum number of users that your AP supports. Then comes the part that most people are not aware of, if somebody is further away from the AP and you are close to it, we are getting into a different model all together.

Even in the simplest scenario where everybody gets their fair share you are most likely never going to be able to see 50Mbps to your device in the near future.
Here in Norway, Netcom (TeliaSonera) have deployed their LTE network and trials by journalists reports that the results they, even at this early stage with presumably few customers on the network, indicates the concurrent number of users will most likely be the biggest issue. 
The best result was 90Mbps during a demo, which probably meant only 1 client was connected to the AP. Then after launch, they saw speeds of 20Mpbs, which is good, but this was in area of the city which probably do not have to many early adapters of tech gadgets, since its mostly an area of traditional businesses and not much housing.
But if you droped into the downtown area you got 8Mbps, these numbers are all download speeds, uploads are of course not even close. Yes, 8Mbps are better than many people have on their DSL links today,  but how does you cloud network business model work with these conditions? 

Okay, so we solve that by installing an AP on each street corner so that every person gets to connect to their own AP, again this does not scale, at least on the economic side. But even if the AP was cheap or people actually payed an extra premium to use one exclusively, then we need to feed the AP with the necessary bandwidth, so lets assume we dont overbook the link, since bandwidth is the key facilitator for the success of the cloud services. Then we need 100Mbps fixed links to the AP's, we still need tons of extra equipment to do it, whether we do it with fiber or copper lines it suddenly has a drastic rate of growth in the expense column.  

So we just went from a FTTH (Fiber to the Home) scenario to a FTEAP (Fiber to every Access Point) and there will be a magnitude of more AP's than there will be homes. And most Service Providers today are not able to get a working business model  in regards to FTTH of the ground, unless they have deep pockets and are investing for the long term future.

Personally I hope that we will get to the 100Mbps per wireless device, but I doubt it will be as easy and come as quick as some others are hoping for. Could a Femtocell type of solution with 4 or 5G support be the answer?

But as the Atheros engineers, I meet at the IPTV Worldforum in London, said; "It all comes down to the physics, we can't do anything about that. just work with it. Heat and attenuation, we salute you Newton!" Not sure why they saluted Newton, might be because gravity makes the beer go down easier?

No comments:

Post a Comment