Sunday, September 6

Fast Forward to 2015: RIP Telcos, Long live brand networks

Was having an interesting lunch today with a friend who founded and we were talking about our intentions to migrate over to google Voice with our cell numbers as soon as google offered it. It was funny, because as we were discussing this, it brought me back to a discussion I had with David Isenberg back when I was running Interoute's IP network.

This was back in 2002 when we were in the last throws of the implosion of the telecommunications industry from the first Web 1.0 cycle, and I had been hired to lead the IP products for Interoute because of my combination of skills in hardware, software and networking. My only exposure to telco land was back in 1995-96 when I built out the first self-service interface for Bell Atlantic (now Verizon) called PICASSO (it stood for something, can not remember the entire acronym).

I built a team, deployed what is now known as application-aware throttling (e.g. net neutrality) and generally got an incredible lay of the land from the point of view of telcos, service providers and our customers. What has become increasingly obvious to me in the past seven years is how easy it is becoming to build networks on top of our old infrastructure, and how the telcos will more than likely become subsidiaries of larger service companies who will build businesses on top of the infrastructure overall.

google's Infrastructure Team
In 2004, I was being interviewed by google for a role to negotiate pricing for dark fiber and colocation space in Latin America. An article by CNET at the time also questioned whether or not google was going into the telco business. A year earlier, I had been speaking with the head of google's Network Infrastructure (who used to be Yahoo's Head of Network Infrastructure) about how to monetize their IP network because of the ease of use of new network technologies and the expansion of the peering nodes that were beginning to be publicly available. Combine the fact that google was a source of more data and could offer it's downstream to other network users that could ride google's network for a much lower cost (since bandwidth prices are all determined by commitment rate).

Additionally, when I ran the IP Product team, we focused on working with the application and hardware providers - speaking with Sony Playstation, Nintendo, and Microsoft XBox about building a private network for their customers to connect to and be assures a specific latency (since that matters in interactive gaming). In every case, the gaming arms of the corporates were not interested in a dedicated network (at that time) since the last mile was the real killer in latency. Instead of being the fastest network, we were hamstrung by the least common denominator (or slowest link in the chain).

Today, google is running it's own global network - a full four years after the start of the effort - more than likely lighting up fiber across the world and colocation centers everywhere. Microsoft has more than likely gotten into the business - and, if AT&T has not hamstrung Apple, they are doing so as well. I know the biggest complaint about the iPhone is not the applications or lack of multi-threading, but the AT&T network not providing the quality of service needed for the phone to operate at its peak.

I know that Sony has dabbled in the business (at Interoute, my team and I built the SoNET for Sony which was a white-label ISP for their camera). So all of the application players are looking for ways of providing network service without dependency on telco service providers.

Arrival of the brand networks
While this might seem silly proclaiming this to be next paradigm (after describing SoNET), I believe that in the next couple of years, we will see a birth of brand networks - IP networks riding on fiber in the gigabit Ethernet and 10G Ethernet range riding on wavelengths of light, instead of the SONET and other optical carrier components that still make up the infrastructure of the large backbone networks of today.

google (if they are able to) will more than likely launch for the first - due to their large bandwidth capability and their value propositions. No one prior to google - or prior to the ubiquity of IP networks - would ever dream of challenging the telco community.

Why? Because the communications networks have been dominated by the POTS network (plain old telephone service) and getting into the physical central office (what the Telecommunications Act of 1986 was supposed to facilitate). Back then, accessing the Central Offices (think of them as the off ramps from the telecommunications superhighway) was supposed to enable new, innovative services across the network.

Now, if you add:
  • with the ubiquity of soft-switches and VoIP - which used to be dominated by real telephone switches in the central offices - at more and more peering points (not owned by telcos),
  • combined with the incredible reach of the Internet Protocol across every network (e.g. mobile networks, DSL, cable, etc),
  • augmented by the network reach through WiFi, WiMAX, EVDO, 3G, 4G and other wireless network protocols and meshing capabilities,
  • accelerated by the ever decreasing cost and complexity of managing world-wide IP networks, and
  • the economics of bandwidth pricing (greater volume means lower prices) from the legacy backbone providers
a brand network becomes even more attractive to companies that rely on them - like google, Apple, Sony, Microsoft, Nintendo and so on.

Why I think I am right
I believe this is going to happen in the next three years because of a couple of things:
  • google Android and Chrome - both are OS platforms for small devices
  • google Voice and number portability - essentially making google the "central office" for all people who switch to google Voice
  • The google Network and the ubiquitous reach it has across the world - which means no other network can be without access to the google network, and
  • AT&T's efforts to crush google in the courts and in Washington DC (as described in this WIRED article)
I would also add the telco's insistence on the need for network neutrality (which essentially means the ability to block packets from specific sources on their network) so they have a tool to beat back google's dominance. The telcos want the ability to selectively filter packets so you would not experience performance degregation when going to your favorite sites - but when a google Voice call is being connected, a performance lag could occur since the google packets would be downgraded in priority for delivery.

When google becomes the network - watch the telcos fall. I assume this will be led by any of the telcos that are generating a significant portion of their revenue by POTS service and have not fully diversified into the IP space. If Apple does the same thing - either through a partnership with google on their fiber or their own purchasing spree - then more networks will begin to fall.

My guestimate on the timeline will be:
  • 2010 - google Voice is on all platforms and networks with number portability
  • 2011 - AT&T pushing very hard to block google's monopolistic tendencies
  • 2012 - google announces google Network - in concert with cable companies and other backbone providers like Sprint
  • 2013 - first of the networks begin to close down (likely T-Mobile in the US)
  • 2014 - google announces deal with Apple to offer Apple iPhone Network for any iPhone on any network
  • 2015 - google purchases Sprint and/or regional providers to manage/maintain the google Network as well as the original POTS/VoIP network
I know my telco friends are going to comment on the last mile/first mile argument - that the telco network that is providing your IP service will always be the weak link in the chain. But, I would argue that google, Palm, Microsoft, Nokia and other handset providers are making that no longer a problem.

On my Palm Pre, I use any wifi network I want to access the world - for uploads, downloads, whatever. I know that when Skype and google put the google Voice on my Pre, I will be using it more than the Phone app - since the Phone app is so much not connected to my life than google Voice or Skype currently is. My Pre will become just another multipurpose computer with a network adapter that accesses any and all networks - and will not be locked to one network or another.

Will this be better than what I have now?
Not sure about this one. I hope so - since google and the other companies are about providing quality of service on their apps - since I would switch in a heartbeat if they did not work for me. Vendor lock is not in their DNA, service loyalty and innovation is.

No comments: