Saturday, July 31, 2010

Backhaul Portion of Harbinger Capital Plan is Where Questions Might be Asked

Of several questions that might be raised about the Harbinger Capital plan to create a new national wholesale fourth-generation mobile network using the Long Term Evolution air interface and satellite backhaul, nothing is more important than the ability to attract enough capital to build the network, and the ability to get at least a few  anchor tenants.

The other issue is how the backhaul can be optimized in terms of latency performance. Putting a fleet of satellites into low-earth orbit is one way to reduce latency. Doing so for birds located in geostationary orbit will be more of a trick. Streaming video won't necessarily be an issue.

But interactive applications, such as gaming, voice, videoconferencing and enterprise apps, might be. The ground-based LTE network is a clever idea. Apparently wholesale customers will be able to lease ground segment separately from satellite backhaul.

But that will pose some issues for would-be wholesale customers as well: how to create the fiber backhaul network supporting all those LTE sites. A rational observer might conclude that a would-be anchor tenant would be better off leasing capacity from Clearwire, which will have the backhaul in place as a necessary part of building out its radio sites, without the satellite latency issues to contend with.

Perhaps there is some clever way to use the proposed Harbinger network's satellite backhaul only for apps where latency is not an issue, keeping latency-sensitive traffic confined to the terrestrial fiber networks. Maybe using it for video on demand only is one such approach.

The issue would seem to be that satellite transponders are optimized for point-to-multipoint video distribution. "On demand" services can be provided, but do not play to the network's advantages.

At some point, complexity becomes the enemy, though. Ways to seamlessly integrate other terrestrial 4G and 3G networks with a separate satellite network potentially are workable. But the cost and complexity might be relatively high.

If demand proves robust enough, one could conceive of a telemetry network build around the Harbinger network that is designed specifically for applications that are latency in-senstive. But Harbinger would not be the first company to attempt a business model for applications and customers that proved not to be sufficient.

http://www.o3bnetworks.com/video/video_03.html?height=315&width=400

No comments:

Many Winners and Losers from Generative AI

Perhaps there is no contradiction between low historical total factor annual productivity gains and high expected generative artificial inte...