O3b Networks LEO Satellite Internet

[sammie]

More information with a focus on Africa, interesting reading

URL updated,

[Ronsmytheiii]

More information with a focus on Africa, interesting reading;

You linked to this thread?

[Danderman]

OK, I must be missing something here. From what I have read, the function of the O3B Networks system is to provide "backbone" or trunk function for Internet in the equatorial regions, ie T1 line bandwidth to rural areas. But, GEO comsats can do this today, albeit with some latency? So, is this $640M project in existence because people in Africa are unhappy with the latency in their current internet connections, or are unwilling to hook up until someone comes along with a quicker system?

[maxx]

OK, I must be missing something here. From what I have read, the function of the O3B Networks system is to provide "backbone" or trunk function for Internet in the equatorial regions, ie T1 line bandwidth to rural areas. But, GEO comsats can do this today, albeit with some latency? So, is this $640M project in existence because people in Africa are unhappy with the latency in their current internet connections, or are unwilling to hook up until someone comes along with a quicker system?

Not only. For example, in January 2008, a very large undersea cable broke and cut off internet access and telecom to a large part of the middle east (see http://www.renesys.com/blog/2008/01/mediterranean_cable_break.shtml  and subsequent for details).
Having alternate routes in case of an incident like this make sense, and it is not rare for large companies to have several ISPs just in case one of them goes offline or experience difficulties.

[pippin]

OK, I must be missing something here. From what I have read, the function of the O3B Networks system is to provide "backbone" or trunk function for Internet in the equatorial regions, ie T1 line bandwidth to rural areas. But, GEO comsats can do this today, albeit with some latency? So, is this $640M project in existence because people in Africa are unhappy with the latency in their current internet connections, or are unwilling to hook up until someone comes along with a quicker system?

Yep. That's exactly what it is about.
Actually that's what pretty much EVERY business case is about: people being unhappy with something.

"some latency" renders a internet connection completely useless for modern applications, today it's much more of an issue than actual bandwidth. You will not be able to use something like Google Mail over a GEO comsat connection, let alone corporate enterprise software integration. I#ve been using a hight latency mobile internet connection for a while, about 1/5th compared to the latencies you get over a satcom connection, but things like this page are a nuisance, more demanding apps are simply not possible.
640 Mil.$ is not a lot of money for that kind of business.
Africa has quite a few people, that's no small potential.

[Danderman]

http://www.o3bnetworks.com/

Some updates on this proposed constellation.

[StuffOfInterest]

Latency reduction could certainly drive a "new model" for Internet trunk access.  New applications such as voice and video conferencing don't like two second ping time.  Things have certainly changed on that front.  In the early 90's, I once had the fun of experiencing a 17 second (yes, 17,000 ms) ping time from a US military site in Asia to a commercial site in South Africa.  Thank you double geo-sync hop!  Having low latency links from places such as Africa to the rest of the world could spawn some significant business development in places such as Africa.

As a side note, when they say "IP communications", I wonder if they mean IPv4 or IPv6?  Someday, the core Internet is going to have to convert and I'd hate to think of a 2/3 of a billion dollar asset stuck without handling the new technology (although IPv6 isn't all that new).

[nomadd22]

Latency reduction could certainly drive a "new model" for Internet trunk access.  New applications such as voice and video conferencing don't like two second ping time.  Things have certainly changed on that front.  In the early 90's, I once had the fun of experiencing a 17 second (yes, 17,000 ms) ping time from a US military site in Asia to a commercial site in South Africa.  Thank you double geo-sync hop!  Having low latency links from places such as Africa to the rest of the world could spawn some significant business development in places such as Africa.

As a side note, when they say "IP communications", I wonder if they mean IPv4 or IPv6?  Someday, the core Internet is going to have to convert and I'd hate to think of a 2/3 of a billion dollar asset stuck without handling the new technology (although IPv6 isn't all that new).

The main reason for IPV6, lack of available IP addresses has eased quite a bit since so many homes and office are on their own networks with cheap routers and aren't eating up public addresses like they feared.
 The goal for satellite internet is straight IP based comms without special interfaces. The most you can push over geo that way is about 400kbs because of packet acknowedgement delay.
 At 7800km they could probably push 2 or 3 mbs over generic IP, and would have to use some sort of special interface for higher.

[drbuzz0]

There have always been really big issues with LEO communications satellite systems which have made them a very difficult place to turn a profit.   Iridium and Globalstar are both magnifficant systems from a technical standpoint - the fact that iridium can provide handset communications reliably to any place on the face of the earth is nothing to discount.   It was and is an amazing engineering achievement.

That being said, the market for their services has turned out to be incapable of making these ventures profitable if you pay for the entire project.   Now, if the company goes bankrupt and it is bought out for cents on the dollar and the debts erased.. well then it can, but that's another issue entirely.

Of course there are some big issues to be tackled here.   One of the biggest is the whole issue of bandwidth.  The LEO satellites like Iridium and Globalstar have really limited bandwidth - no more than 9.6 kbps or so.   That's hardly useful for anything beyond text communications.  Voice on Iridium is extremely compressed using AMBE, one of the most advanced voice coding formats around to save bandwidth.

The issue has to do with the frequencies used and the types of antennas used.   If you have a fixed antenna and a geostationary satellite then you can use a point to point beam and make use of bands like the Ku and Ka band which have a lot of bandwidth.  The antennas are dish or phased array types with a narrow beam so you don't have to share the frequencies with everyone.  Move the dish a few degrees and you can talk to a different satellite on the same band.

Of course the LEO satellites don't stay in one place and are generally used for mobile stuff.  For that reason they use the l-band and the devices have omnidirectional antennas.  They can't effectively use the higher bands and they can't use narrow beams because you'd have to track the satellites constantly.

Of course, this is not impossible if you have a fully steerable set of antenna arrays - you could track each satellite as it comes into and out of view and then pick up the next one.   This is necessarily large and needs to be stable and have a good view of the sky so it precludes handsets or small devices like that.

The GEO systems have been more successful, for both mobile and fixed sights.   Only one satellite is needed to cover an entire region, you can use higher bands relatively easily and the satellite can be multi-purpose.   

There are mobile GEO systems as well.  Some of them use small dishes or phased arrays that are poined at the satellite either manually or, in the case of a ship or plane, automatically by an in-motion tracking system.  These are such systems as Inmarsat and Msat.   THere are a few next-generation ones like Thuraya which can actually make use of handsets.   Thuraya does this by using a massive deployable high gain antenna array on the satellite which gives it the gain to connect with small low power transmitters on earth even at such an altitude.

There are also the GEO services that offer true broadband and concentrate on fixed sites.  These are ones like Hughesnet (formerly Direcway, formerly DirecPC), Starband and WildBlue.   They use the KA or Ku band and have been commercially successful to a limited degree.


The issue with these, in addition to having to place the antenna pointed at the satellite and having an unobstructed view of it is latency.   The latency of satellite internet is bad, but it's not absolutely intolerable.   It depends on the activity.   Using a GEO satellite for something like telephone is totally doable.  The latency is a minor annoyance but it's not a deal breaker.   Using GEO for surfing the web is very usable.   If the system is properly setup and uses techniques like prefetching and some caching then the latency is hardly noticable.

Of course for things like downloads and one-way streaming it's not an issue.   But it does make teleconfrencing haphazard and things like gaming impossible.

The question is:  is it worth the enormous costs of a LEO satellite system to provide latency-free internet?    The GEO systems work, they work decently and they have their issues, but it's really a BIG jump to get these resolved. 

[pippin]

Of course the LEO satellites don't stay in one place and are generally used for mobile stuff.  For that reason they use the l-band and the devices have omnidirectional antennas.  They can't effectively use the higher bands and they can't use narrow beams because you'd have to track the satellites constantly. 

My understanding is that this is NOT the case here. The idea is to provide backbone service, not direct access so you will not have to communicate to thousands of handsets at a time and the antennae don't have to be omnidirectional, at least not the ones on the ground.
Also power (on the handset side) is not an issue with this approach, which has been the biggest problem for Iridium.

[agman25]

Some updates...

http://www.spacenews.com/satellite_telecom/o3b-networks-ceo-expects-close-billion-financing-package-soon.html

[bolun]

O3b 'fibre in the sky' project wins financial backing

The financing has been secured for one of the most ambitious commercial space projects of the decade.

O3b Networks has raised $1.2bn (£700m) for a series of satellites to support super-fast broadband connections to Africa and other emerging markets.

The company informed the markets early on Monday that a collection of investors and banks would provide the money it needed to launch the venture's first eight satellites.

These will be constructed by Thales Alenia Space at its manufacturing facility in Cannes, with the first platforms ready to go into orbit in the first half of 2013.

Russian Soyuz rockets will launch the satellites from the new Sinamary spaceport in French Guiana.

http://www.bbc.co.uk/news/science-environment-11847708

[zaitcev]

This almost sounds reasonable. Without sat-sat links it's basically another Globalstar, only with one orbital plane and a promise of better data rates. The question is, who's going to buy the service.

[Nate_Trost]

I suspect the targeted customers aren't individuals with phones, but small mobile services operators with remote cellular towers and very limited national infrastructure.

[Robotbeat]

I suspect the targeted customers aren't individuals with phones, but small mobile services operators with remote cellular towers and very limited national infrastructure.

I believe LEO-based sat phones will eventually start crowding out cell phone providers in rural areas (especially). Once you have the infrastructure for a very broad sat phone service (with LOTS of active calls per satellite) in place (and go bankrupt so you don't have to pay for it anymore... lol), then you can compete in almost every cellphone market in the world at a far, far lower fixed cost than if you wanted to do the same thing with cell towers all over the world. It's like a shortcut to global domination. If you can pull off making it cost-competitive with cellphones, then it would be immensely profitable. If you survive long enough, that is.

[kevin-rf]

Hasn't happened yet, they need to get the price point and quality to better than the cell providers. It places an unnecessary cost burden on the end users. Plus once a tower goes up, it is up. Fiber had a similar effect on the GEO market.

Now that anyone with an internet connection can throw up a "micro tower", the necessary dead zones can be well covered (Actually threw a Verizon network extender up in the house to cover "my" dead zone)...

[wjbarnett]

Plus, the cost of sat phone components will always be higher than that traditional cell phone alone; then add the difficulties of uplink RF signal penetrating building materials (sure, rural mostly-wooden building less problematic than metal; but lots of metal roofs where I am).

I've done the trade studies (professionally): won't happen! except for niche markets, or possibly in a collaboration/hybrid ATC (Ancillary Terrestrial Component) deployments, like TerreStar and their traditional wireless network partners.

[Robotbeat]

Hasn't happened yet, they need to get the price point and quality to better than the cell providers. It places an unnecessary cost burden on the end users. Plus once a tower goes up, it is up. Fiber had a similar effect on the GEO market.

Now that anyone with an internet connection can throw up a "micro tower", the necessary dead zones can be well covered (Actually threw a Verizon network extender up in the house to cover "my" dead zone)...

Actually, maintaining a tower isn't cheap.

And, yes, being able to penetrate into a building is a quite significant problem for sat phone providers. If then can overcome this limitation, they will have a chance at doing what I described.

If you can get a million subscribers, it doesn't take actually that much more orbiting infrastructure to do 10 million or more... But you have to reach a critical point where you have enough subscribers to make it cheaper than cellphones are. At that point, relatively small investments in capacity can result in vastly expanded service.

This is dependent on technological advancement. You'll need a bigger, more powerful satellite to be able to service almost a million calls through building materials. Larger solar array and larger antennae both help. This is many orders of magnitude more powerful than current Iridium satellites.

[wjbarnett]

"power" on the sat side is just a part of the problem. As the sat or any single RF source gets more powerful, the much bigger problem becomes interference. So if there are many subscribers near each other (in sat terms, is within the radius of the same spot beam -- ie 50 to 100s of miles; vs cell tower of <5) will cause problems! Generally n=64 simultaneous bearer paths is the max supported in 3G/WCDMA on the same RF carrier within the same sector/cell. (I'd have to lookup the value in LTE/4G; but much improved due to ODMA). But still ODMA is approaching the Shannon limit for information transmissions (considering real-world system losses) and therefore has required growing the RF carrier size to 10 or even 20Mhz (one-way). Sats won't have enough spot beams nor their operators enough spectrum allocated/purchased.  Sure 10m subs, no problem. 3B subs = fantasy.

[jongoff]

Plus, the cost of sat phone components will always be higher than that traditional cell phone alone; then add the difficulties of uplink RF signal penetrating building materials (sure, rural mostly-wooden building less problematic than metal; but lots of metal roofs where I am).

I've done the trade studies (professionally): won't happen! except for niche markets, or possibly in a collaboration/hybrid ATC (Ancillary Terrestrial Component) deployments, like TerreStar and their traditional wireless network partners.

I like the idea of ATCs (thanks! I didn't know the technical term).  If you can't beat 'em, join 'em.  :-)

Do you see any possibility of LEO-based ATC satphone/cellphone combos making much headway into the market?

~Jon