Innovation is also about evolution, and not always revolution. Sure, if money doesn't matter it would be wonderful to put up a new constellation of LEO satellites with built in routers to allow any remote site to communicate directly with any other remote site over a single satellite hop.

The problem with this approach is that it doesn't start to work until the entire constellation of new LEO satellites is rolled out - at tremendous expense. A LEO satellite doesn't stay overhead, so connections must be passed from bird to bird as they pass overhead. The customer equipment is more expensive as it must track the birds as they go by.

If you started off by installing router technology in GEO satellites you could immediately enjoy the benefits on the entire footprint of that satellite. Anyone within that footprint would be able to get site-to-site connectivity without going through a central hub.

Aside from the reliability/availability issues, you also need a new service model. Most satellite providers want to sell space segment to network operators. They don't want to be the network operator themselves - delivering services to end customers and handling NOC/TAC support calls. Their business model is based on selling chunks of space segment to network operators who then deliver services to the end customer, much as fiber backbone providers provide access to their backbones to service providers.

Technology has continued to evolve, and very rapidly so in the satellite industry. Just 5 or 6 years ago very few broadband satellite solutions had TCP Acceleration technology. Today most of them do. Two way commercial broadband satellite service is actually pretty new, and companies such as iDirect have improved the efficiency dramatically over older Aloha-based bandwidth allocation mechanisms. Today you can get multi-frequency, deterministic TDMA on the inroute (upload) that is highly efficient and supports real-time applications such as jitter-free VoIP, Quality of Service and cRTP header compression. DVB-S2/ACM technology is bringing significant increases in efficiency to the outbound (download) carriers, and providing mechanisms to change modulation coding on the fly as rain fade conditions come and go.

I think you have to balance the realities of the business model with the practicality of new technological solutions. A new generation of routed, single-hop LEO satellites sounds wonderful - but is it realistic in the foreseeable future?

The article in question started off..."In this industry, there's a lot of talk about branch offices and satellite sites." Perhaps it was my misconception, but I interpreted "this industry" to mean the broadband industry, and in particular the broadband satellite industry. The authors were apparently unaware that technology has evolved with respect to the latency issues that confront TCP, but the concept of adding routers to satellites as a further evolutionary step is quite interesting. However I believe there are availability/reliability issues that need to be solved first in order to take that step. I expect it will happen eventually. It really means putting a "hub" on the satellite.

Evolution, not revolution. It's always been that way. The broadband satellite industry is going through the same thing the terrestrial industry did in the 90's as everything migrated from point-to-point leased line (9002) circuits to Frame Relay, ATM, SMDS, ISDN, etc. and these technologies fought it out in the market for dominance. In the last few years we saw iDirect's proprietary TDM/MF-D-TDMA technology compete with standards based DVB/RCS technology. (iDirect's technology clearly won the enterprise market, while RCS struggles on in the residential/SOHO market). It's been an exciting time in the satellite industry these last 5 - 7 years; but few people outside the industry are aware of it.

Patrick Gannon
Business Satellite Solutions, LLC
www.bsatellite.com