Two-Way Satellite Internet Is Here!

spectro writes: "Wired is reporting the first two-way satellite Internet service has been launched by StarBand. The service promises speeds up to 500Kbps down and 150Kbps up, but a ping latency of about 400ms, so gamers are out of the question. Anyway a nice alternative for those of us who cannot get DSL yet, but watch out... The Evil Empire is part of the joint venture." It's nice to know that someone has finally made the leap, after years of promises and millions of R&D dollars. Check out the article for information on some of the competition, too.

Re:"very low pings"

The numbers are very easy to calculate. 22,000 miles, times 4, divided by 186,000 miles per second; that is 473 mS just for time of flight.

The service should shine for big downloads, but be rather poor for highly "chatty" stuff, with many request-response pairs. Loading web pages with many small objects would be rather disappointing, I should think. Checking POP3 email, if there are many small messages, would be pretty poor.

It's been here for a while

It's just that it's not that commonly used. I installed a bidirectional, mobile satellite internet link 2 years ago, with T-1 speeds. Granted, not the cheapest thing to do, but it was still there. And it sucked, at least for web surfing, which was why we installed it. Unfortunately, we didn't have much of a choice since we needed a mobile solution, and Ricochet wasn't up to speed yet.

Typically, satellite data streams have about a half second latency. That means you send out a ping to your next hop, and it's at least 500ms before you get the response. Now when you're doing a stream of data in, this isn't so bad. You have a half second latency when you're setting up the connection, but then you have a nice even data delivery. However, for things like web surfing where you're setting up lots of connections (up to 30 or 40 per page sometimes), it's unbearable.

Satellite internet connections are useful if you're in a remote area and can't get anything else. But if you have anything else, even IDSL, you're going to find it a better choice for things like web surfing and gaming.

-Todd

---

Re:It's been here for a while

It probably isn't as bad as you suggest. You shouldn't be seeing 30 to 40 connections to pages with modern (or at least future) browsers and servers. After getting the HTML, the popular browsers usually open up 4 to 7 concurrent connections which incur the round trip times in parallel and then reuse them for subsequent requests when possible.

I'm sorry, perhaps I wasn't clear when I said that I installed and used this setup. The particular site we were using was washingtonpost.com, which at the time averaged 50 objects per page. It wasn't the worst site out there (ICQ was much worse), but it definitely wasn't the smallest pagesize.

Even if you have 4 simultaneous connections and reuse them, you're still incurring large delays. To set up the TCP connection, you need a three way handshake. That is a minimum of 1.25 seconds (given .5 seconds for a round trip). Then, best case scenario, you send a request and get a response for another 1 second of latency, plus any latency present between the satellite headend and the website. Let's assume 40 objects on the page. So each connection has to get 10 objects in parallel with the other connections. So that's 1.25 seconds to set up the connection, and 1 second per image. That's 11.25 seconds of latency per page added by the satellite connection. Theoretically, if you send all the requests at the beginning of the connection, you can reduce that latency. However you still have a minimum added latency of 2.25 seconds, and it's questionable whether or not you can do that with current browsers (I personally don't know).

In contrast, if you open up a dialup connection to an ISP without caching or compression and pull washingtonpost.com's front page, it will take about 11 seconds to download and render the page. That number is based on a series of tests that I performed personally, using several different computers to make sure the number was accurate.

Also, if you're wondering where I got a half second for a round trip from, it's fairly simple. The satellite is located 22,500 miles above the equator. Given a best-case scenario with both you and the satellite headend on the equator, that's a one way trip from you to the headend of 45k miles, and a RT of 90k miles. The speed of light is 186,300 miles per second. So that gives a best-case latency of about a half a second.

-Todd

---

Shit...

Why not just use AOL?! 2k up and down is all anyone really needs. 2000 bytes is alot.

Re:What about mobile use....

This is very empowing stuff.

And that's not the half of it. Imagine you're a foreign newscaster stationed in the backroads of Afghanastan or Somalia. Do you know how much your life is put in danger every time you try to "link up" and communicate with the agencies back at home? Most often, the telephone infrastructure is nonexistant, and the rest of the time, it's bugged.

Or imagine you're the driver of a van delivering groceries and blankets to the homeless. Do you know how many lives are lost every year owing to exposure, simply because of miscommunication between relief agencies and dispatchers? A continuous link with home would solve that dilemma while providing incidental benefits like letting the homeless check their email or search for jobs on the internet.

Radio will get you somewhere, but cbs are subject to a lot of abuse. Recently in NY, disgruntled ambulance drivers were (illegally) jamming the airwaves by blowing on the receiver each time a dispatch went out to a non-union ambulance. Thankfully, no one was (apparently) killed by the practice, but just the same, it's a scary prospect, and it'd be a lot harder to jam a satellite feed.

The sooner we realize real lives are at stake, the sooner we'll embrace this technology, for the greater good of humanity.

Buzzword Security

From the article:

The DirecPC system uses 56-bit encryption on the packets of data that are sent downstream from the satellite, according to Steven Salamoff, assistant vice president of the DirecPC business services division. Deciphering upstream packets would first require breaking the downstream encryption because of its random packet generation, Salamoff said.

Consumer awareness of home computer security is growing. And it looks like the DirecPC salesforce is chomping at the bit to try out their service's embeded buzzword: encryption. Yep. They're safer than that Cable and ADSL you hear about. Unlike those guys, satalite providers encrypt their data.

Sure, wireless brings forth a whole new series of security concerns. And that means encryption has its part. But it does nothing to address insecure hosts on persistant broadband connections.

What do you want to bet that the new satalite service providers will do the same as their xDSL and cable competitors and ignore this problem. But hey... they have encryption.

Geostationary Broadband Services

In late 1991, as vice president for public affairs at E'Prime Aerospace I did a lot of the work to break up the log-jam within the FCC on the first Ka-band satellite license ever to be issued. It was for a two-way geostationary broadband system called "Norstar" that was based on Milstar satellite technology. The way these geostationary systems work is to use a phased array antanna on the satellite to rapidly multiplex a tight spot beam between ground stations and dynamically shift the time slices to various geographies based on load requirements.

The higher the frequency allocated for the service, the tighter the spot beams with a given dish diameter, which means you can end up with the geographic equivalent of a dynamically distributed cellular communication system deployed via a single satellite with the main drawback being that you are sharing a single point of failure as well as a bandwidth bottleneck. However, at these high frequencies satellites can be parked much closer to each other in the same orbital so the bottleneck and failure vulnerability can be minimized by requiring a bit more complicated ground station to allow a fixed dish to rapidly switch between co-located satellites.

The rain problem is serious, but can be minimized, at least on the downward leg, by increasing the energy storage capacity of the satellite to power through the weather. IIRC the upward leg has a bit more of a problem with heavy weather because, although power is quite available on the ground, the scattering occurs far from the satellite (clouds are only a few miles in altitude whereas the satellites are about 20 thousand miles away). On the upward leg it is a bit more problematic due to the fact that arc very tightly which further increases the density of communication via demultiplexed (parallel) communication channels.

The delay time introduced by going up and back is less than typical human reaction time (about a quarter second) so its just fine for most practical uses -- even interactive ones other than games where people are pushing human reaction times to the sub 100ms ranges.

Starband is probably in these high frequency ranges since they do have a bit of difficulty with heavy weather, although I haven't been able to locate their FCC filing online to see exactly how high the frequency actually is.

one pipe- bandwidth limitations

While the article doesn't come out and say it, I'm guessing there will be serious total bandwidth limitations- since there are no tracking antennas, and since this is referred to as part of a satellite TV service, it must be from (and probably to) a satellite in geosync orbit. All traffic is probably going to just one satellite. Once a lot of people sign up for this- popularity will kill it- since I'm sure that they can sign people up to long term contracts faster than they can put more transponders up.

Plus there is the fact that anything beamed to you is probably also being beamed to everyone in a multiple state area around you. I sure hope they have some *strong* encryption built in. They ought to be doing something better than DES, no?

I hate to say it, but the microsoft backed Teledesic system is a much better system (many satellites in low earth orbit- kinda like a cellular system) the satellites are closer and cover smaller areas, so the amount of bandwidth/satellite is much smaller.

Whatever happened to a simple net connection?

Hardly anyone wants to sell you a simple net connection anymore. What's with all the useless addons? ISPs are always trying to be their own portal when they can't top Yahoo. Best Buy tries to get you on MSN when you buy a refrigerator.

Now to get satellite net access you have to buy a whole new PC, and it's still $60 a month?

Dammit, give me some hardware and tell me what I should set the IP and DNS to, and charge me less!

Damn light is so slow

but a ping latency of about 400ms so gamers are out of the question.

I rember when i thought light (186,000 miles/second) was fast haha. Now 400ms a second isn't round trip time for light from nowheresville -> Satellite -> Backbone router but the handicap makes up for a damn sizeable chunk

Let's investigate.

For example, if the mission requires a geostationary orbit, which can be achieved only at a distance of about 35,000 km (22,000 mi) above Earth.

I got this from MSN Encarta. MSN... like their operating systems, probably can't be trusted :).
Lets do some calculations(my favorite). For light to get to the satellite it takes .118 (approx) seconds now multiply by 2, we have to go both ways. we have .236 seconds (approx).
Now a handicap of 236ms before you get on the net leaves a little to be desired. Like, Damn I wish light were faster.

Time is Change