Monday, March 11, 2024

Starlink has begun delivering promised latency cuts

In his January 12th  SpaceX update, Elon Musk said the biggest goal for Starlink from a technical standpoint is to get the mean latency below 20 ms. He expanded by saying that given the speed of light, 8 ms is the absolute minimum latency for a satellite at 550 km. He believes they can optimize terrestrial and inter-satellite links, and minimize queueing delays and dropped packets, to recude the the rest of the time to below 10 ms. He predicted that eventually"Starlink will be more responsive than ground Intenet in most cases."

A month later, we saw early results of the latency-reduction effort. On February 12, Oleg Kutkov tested Starlink's Rev 3 and Rev 4 terminals and, as shown here, he found no latency inflation as background upload and download speeds increased simultaneously when using the Rev 3 terminal. He compared the Rev 3 and 4 terminals and found that Rev 4 upload and download speeds were about 50% faster than Rev3. Average ping times were somewhat improved for Rev 4 (88 vs 93 ms), but jitter was significantly lower (9.2 vs 111.9 ms).
You should also check out Dave Taht's take on Oleg's results. He concedes that Starlink has improved dramatically, and outlines steps they could take to further reduce latency.

Last week, SpaceX released news of progress toward the 20 ms latency goal. They have worked to reduce latency throughout the Starlink system. Since the begining of the year, they have deployed and tested 193 different satellite software builds, 75 gateway software builds, 222 Starlink software builds, and 57 WiFi software builds. 

This is the latency view of the interactive map.
For a month preceding March 7, SpaceX collected data every 15 seconds from millions of Starlink routers. In analyzing the data, they defined worst-case latency is the point at which 99% of times are shorter and peak hours as 6-9 PM local time. 

In the United States, they found that median latency was reduced by more than 30%, from 48.5ms to 33ms during peak usage hours, and worst-case peak hour latency had dropped by over 60%, from over 150ms to less than 65ms. Outside the United States median latency was reduced by up to 25% and worst-case latencies by up to 35%. The map shown here is interactive and shows availability and upload and download speeds in addition to latency.

SpaceX says it has “tuned our algorithms to prefer paths with lower latency, no matter how small the difference, and to remove any and all sources of unnecessary and non-physical latency." Dave Taht and his colleagues at  LibreQoS might disagree, but latency will improve over time regardless. 
Latency will improve as SpaceX launches more satellites with more capacity and inter-satellite laser links and the launch rate will increase when Starship becomes available. Adding ground stations will also improve latency.  (Note that the only African light-colored areas in the above latency map are within reach of the only ground stations on the continent).
I hope Musk achieves the 20 ms goal for Starlink. Doing so would not only benefit Starlink customers, it would call the attention of the FCC and terrestrial Internet service providers to the importance of latency as a performance and marketing metric.

Thursday, February 22, 2024

Civilian Tech Mobilization in Ukraine

Rosie the Riveter, US World War II poster (source)
As was the case in the US during World War II, civilian volunteers are making important contributions to the Ukrainian war effort.

On February 8, 2022, the first truck load load of Starlink terminals arrived in Kyiv. A week later they were being used. By April 2022, there were 5,000 terminals in Ukraine, and 42,000 as of April 2023. (At this point, SpaceX and Ukraine have gone silent. Neither ChatGPT4, Gemini, Copilot, Perplexity, nor I could not find a current terminal count).

Whatever the number of terminals in the country, they require support. They were purchased, delivered, and set up. Users were trained and they require real-time access for troubleshooting and assistance. Broken terminals have to be repaired and some terminals have been modified. Civilian tech volunteers are doing much of this work. 

There are several Starlink support centers throughout Ukraine. For obvious reasons, they are secretive about their work, but one large one is Nebogray in Lviv. Neborgray has repaired 5,976 Starlink terminals and converted 516 for portable use mounted on vehicle roofs. In addition to the service centers, there are many individual craftsmen and small services throughout the country.

The work at Nebogry is performed by highly qualified volunteers. For example, Oleg Kutkov. is a senior engineer at Ubiquiti, and he devotes his spare time to Starlink research. He bought what may have been the first Starlink terminal in Ukraine on eBay before the war and does teardowns and research studies like this recent unboxing and review of the Version 4 Starlink terminal on his blog. Oleg is an active participant in the Starlink mailing list and the 15,700-member People's Starlink Facebook group.

The Facebook group was created by The People's Starlink project, which is involved in refurbishing, adapting, repairing, and providing technical support, as well as procuring and upgrading satellite communication terminals from SpaceX's Starlink for the needs of the Armed Forces of Ukraine and other defenders of Ukraine. 

With the help of many contributors, including Oleg, People's Starlink founder Vladimir Stepanets has written a 246-page Starlink Handbook for Military Users, which begins with a message from the author “Greetings defenders of Ukraine!”

Starlink Handbook for Military Users
The handbook is divided into eight, richly illustrated modules:

  • What is Starlink?
  • Starlink Terminals
  • Powering Starlink terminals
  • Expanding and collapsing Starlink terminals
  • Terminal management and settings
  • Safety of using Starlink terminals
  • Diagnostics and problem-solving
  • Starlink in network infrastructures

This is the second (and first public) edition of the handbook and it will continue evolving. It is currently available in Ukrainian, but Stepanets is discussing translations into several other languages and plans to publish it as a book.

Point of invincibility in Bucha, Ukraine
In response to Russian attacks on critical infrastructure, Ukraine has established thousands of Points of Invincibility, tent structures equipped with generators. The government is working to provide a Starlink terminal for each of them in addition to heat, water, lighting, and more.

The IT Army of Ukraine is an international, loosely connected organization of Ukrainian and foreign ethical hackers. They have created an online service that Ukrainian allies can use to generate denial-of-service attacks. Of course, one man's "ethical hacker" is another man's "terrorist," and Ukraine has petitioned The International Criminal Court to investigate Russian cyberattacks as war crimes. The International Committee of the Red Cross has published rules of engagement for civilian hackers involved in conflicts and the IT Army will make a best effort to follow the rules.

It was obvious from the early days of the war that two technologies -- Starlink and drones -- were going to play major roles. Model airplane hobbyists created an air reconnaissance unit within the army when fighting began in 2014 and Starlink enabled surveillance drones to relay target coordinates to artillery units. 

Today, non-technical people like Violetta Oliynyk, an artist and jeweler, are assembling drones in their spare time. She learned drone assembly by taking an online course from Prometheus, a nine-year-old education site with over 400 courses online. (The course was developed for the Victory Drones project). Social Drones UA is another volunteer drone assembly project. They vet then train and support potential assemblers with a how-to video and online support. 

Ukrainians are also assembling battery packs from batteries in discarded vapes, which is reminiscent of Americans saving and turning in excess cooking fat to be used in explosives during World War II.

Civilian volunteers and Ukrainian tech companies have pivoted to military innovation and production. Ukraine was technologically advanced before the war and has been forced to innovate and improvise. If Ukraine survives, the tech sector will thrive when peace comes. 

I've presented a few examples of civilian tech support for the Ukrainian resistance to the Russian invasion. There are many others, and if you are so inclined, the Internet makes it possible for you to contribute to them. Many project Websites have contribution links and you can also consult Reddit's list of vetted Ukrainen charities.

Update 3/8/2024

I believe Oleg is by Volodymyr Zelenskyy's ear.
Oleg Kutkov was among the "people embodying the spirit of Ukraine," chosen by Time Magazine as Person of the Year for 2022. Time wrote:
Ukraine first came back online when Elon Musk activated his low-­altitude Starlink satellite internet, as he would later do in Iran. The net was crucial to Ukrainian forces, who were issued the compact, portable Starlink antennas. But in Kyiv, self-­described “tech and space nerd” Oleg Kutkov reconstructed a Starlink dish from eBay, and after contacting SpaceX support, caught a signal. “I was the first civilian user of Starlink here in Ukraine,” says Kutkov, 34, who began a Face­book group that has grown to 8,700 people. “They read about me in the news, and they were all worrying about connectivity because the internet is really important here to get all the news, to get notifications and so on.”



 


Friday, January 19, 2024

Amazon Project Kuiper vs SpaceX Starlink

Amazon's Project Kuiper is far behind Starlink and is under time pressure, but Amazon has several things going for it.

In 2019, I wrote that Amazon would be a formidable satellite-ISP competitor. I still think so, but I didn’t expect it would be over four years until they launched the first test satellites. In the meantime, SpaceX has put over 5,000 satellites in orbit and has over two million Starlink customers.

Amazon has permission to launch 3,236 satellites. They must manufacture and launch at least half of them by July 2026 and the remainder by July 2029. Can they do it? After many delays, they have finally launched two test satellites, confirming that inter-satellite laser links (ISLLs) worked at 100 Gbps while sending traffic “in both directions from the internet over an AWS fiber-optic connection to our ground gateway station, up to our satellites, and then down to a customer terminal at our test location.” All Starlink satellites launched since September 2021 will have ISLLs, so by the time Kuiper is complete, July 2026, all or nearly all Starlink sats will have them and they will have a much larger constellation.

Amazon has not launched any production satellites and they will have to hurry to meet the 2026 and 2029 deadlines. They have signed contracts for 83 launches over a five-year period, which they say will provide capacity for “the majority” of the constellation. SpaceX was conspicuously not one of the vendors, and a shareholder lawsuit pointed out that Amazon had not considered SpaceX as a provider and nearly 45% of the overall value is for launches and engines from Blue Origin, a rocket company founded by Jeff Bezos. Subsequently, Kuiper signed a 3-launch contract with SpaceX.

(Note that Blue Origin has not yet launched their forthcoming New Glenn rocket, which was initially scheduled to fly in 2020. The New Glenn will have greater capacity than SpaceX’s current Falcon rocket but significantly less than their forthcoming Starship).

Amazon's Project Kuiper is far behind Starlink and is under time pressure, but Amazon has several things going for it:

  • In his first letter to stockholders, Jeff Bezos stressed that Amazon was an infrastructure company and that has been borne out by subsequent investments in facilities and services. Amazon will bundle Kuiper access with data storage and cloud computing services.
  • Kuiper will offer service-level agreements to non-consumer customers.
  • Amazon will be Kuiper’s largest customer. With over 1.5 million employees staffing offices, warehouses, and other facilities, Amazon will use the Kuiper constellation internally as will their fleets of delivery trucks, planes, shipping containers, and perhaps delivery drones someday.
  • Amazon is already in the space business with its satellite ground station service.
  • The US, Taiwanese, and other governments and militaries will see Amazon as a more reliable supplier of critical infrastructure than Starlink given Elon Musk’s political activism and Tesla’s dependence on China.
  • Some potential customers may not approve of Elon Musk’s political involvement.
  • Re-usability gives Starlink a large launch-cost advantage, but if Elon Musk can afford Twitter, Jeff Bezos can afford Kuiper.
  • Amazon's New Glenn rocket is designed to be reusable and eight Chinese private and state-owned entities are developing reusable rockets.
  • Kuiper will be launching state-of-the-art satellites and selling state-of-the-art terminals.
  • Satellite antennas are expensive, and Amazon has experience designing and manufacturing consumer devices like the Echo and Kindle. 
  • Amazon has announced three Kuiper antennas.
  • A user terminal is more than just an antenna and Dave Täht, Chief Science Officer at LibreQoS, has been calling attention to Starlink’s latency problem for years. It now seems Elon Musk is ready to listen. Amazon should talk with Täht.
  • Amazon is already talking with enterprises, governments, schools, hospitals, and mobile operators. They have contracts with Verizon, Vodafone, and NTT and licenses to operate in more than 15 countries, including Brazil, Canada, France, Mexico, and the US.
I could go on but you get the idea -- I think Kuiper will survive despite a rocky start and will eventually offer Starlink healthy competition.

Update Feb 9, 2024

I listed Jeff Bezos’s wealth and the expected use of the Kuiper constellation among the causes of my optimism. A report on the capital expenditures by the three major cloud service companies puts Amazon’s commitment to invest $10 billion over several years in Project Kuiper in context. In 2023 alone “Amazon’s (relative) CAPEX austerity continues, as the company spent a measly $53.7 billion, a decline of 20%" and it has invested around $380 billion since 2000.

Monday, December 11, 2023

It's the latency, FCC

Section 706 of the Telecommunications Act of 1996 orders the FCC to “encourage the deployment on a reasonable and timely basis of advanced telecommunications capability to all Americans.” On October 25, The FCC issued a notice of inquiry (NOI) into how well we are doing and invited comments.

The NOI points out that COVID and the concomitant increase in the use of interactive applications has “made it clear that broadband is no longer a luxury but a necessity that will only become more important with time" and proposes “an increase from the existing fixed broadband speed benchmark of 25 Mbps download and 3 Mbps upload (25/3 Mbps) to 100/20 Mbps." They also seek comment on a long-term speed goal of 1,000/500 Mbps.

The focus is clearly on speed. They mention latency on page 12 and jitter and packet loss on page 15, but the FCC made no metrics recommendations on those metrics and requested comments.

Dave Taht, Chief Science Officer of LibreQOS and an embedded Linux developer and consultant since 1998, drafted a comment arguing that the FCC should “balance its near-term efforts on achieving Internet resilience and minimizing latency, instead of only increasing speed.” Taht invited experts to suggest edits to and sign his draft and the submitted comment has 63 signatures, many of which would be familiar to CircleID readers. 

Taht says “Calls for further bandwidth increases are analogous to calling for cars to have top speeds of 100, 500, or 1000 miles per hour” and the "only way to improve responsiveness is to robustly and reliably reduce the latency, and especially the 'latency under load'.” He points out that low latency, not speed, is critical for today’s interactive applications and high latency reduces aggregate network efficiency and increases variability in the user experience. 

Much Internet latency is caused by bufferbloat – packets working their way through queues that build up in routers and other network equipment. Taht has spent years developing tools to measure latency and reduce bufferbloat and he documents his work and that of others in his 27-page NOI comment.

How much speed does one need? 

That depends on the applications you use, which is a moving target. My first home Internet terminal was a 10-character per-second (CPS) ASR-33 Teletype with an acoustic coupler. I used it for email, FTP, Telnet, and network news and I was able to collaborate with distant colleagues. I loved it and 100 CPS would not have made a big difference because 10 CPS was about as fast as I could read and faster than I could type. My first connected computer used a 300-bps modem, and modem speeds increased to 56 Kbps driven by applications like Web, and voice over IP.

Today, Poa Internet in Kenya offers uncapped 4 Mbps service which is sufficient for downloading software, articles, books, movies, etc., shopping, making voice-over-IP calls, listening to podcasts, reading newspapers, etc., and, importantly, creating content and inventing and developing applications and services that are relevant to Africa.  

Netflix speed recommendations (source)
Streaming video is the most speed-intensive application I use today and Netflix recommends 15 Mbps for viewing UHD 4k movies. Poa Internet customers might be able to view 720p video.

Spectrum, my ISP today,  offers three plans – up to 300, 500, and 1,000 Mb/s.  I have a 300 Mbps cable connection which is more than I need. M-Lab's Internet performance test service, which measures speed and latency unloaded and while simulating background activity reported that my latency increased from 16 to 53 ms when downloading was active and 41 ms when uploading was active. Speeds were 355.3 Mbps download and 11.2 Mbps upload. Considering Netflix’s recommendation, it is unsurprising that streaming two movies on my home WiFi network while running the M-Lab test did not make much difference.

As long as I only watch one movie at a time, I suspect I would not notice much difference if Spectrum only provided me with the current FCC benchmark of 25/3 Mbps. This raises the question of opportunity cost. How much capital and operating cost could Spectrum have saved if they had only provisioned, say, a choice between 25/3 and 50/6 Mbps? Would the savings be sufficient to fill in white spaces in their national broadband map

Spectrum dismisses latency, writing:

Latency is typically measured in milliseconds, and generally has no significant impact on typical everyday internet usage. As latency varies based on any number of factors, most importantly the distance between a customer's internet connected device and the ultimate internet destination (as well as the number, variety, and quality of networks your packets cross), it is not possible to provide customers with a single figure that will define latency as part of a user experience.

Page load time as a function of latency (source)
If we could come up with a "single figure" to define and measure latency, ISPs would have an incentive to improve it, and the FCC could adopt benchmarks. While a single figure may be impossible, could tests isolate the latency in an ISP network and the customer premises equipment (CPE) they supply? Could we use imperfect surrogates for latency like page-load times? Could we benchmark components like the CPE an ISP provides?

While the FCC and ISP marketing are focused on speed today, attention to latency and its measurement is growing within the technical community. To learn more and get involved, check Dave's Bufferbloat.net site and LibreQOS and watch Dave's talk here. You can also give the FCC feedback by commenting on Proceeding 22-270 on the FCC Express Comments Page.

Update 1/18/2024

Elon Musk summarized SpaceX's 2023 accomplishments in a recent talk at Starbase in Texas, He covered many topics including Starlink. He stated that their biggest single technical goal for the year was to get mean latency under 20 ms. (He estimated that 10 ms was the theoretical minimum given the speed of light). Doing so will require a combination of steps including launching satellites with inter-satellite laser links, adding ground stations, and heeding the advice Dave Taht has been offering for years.

Update 2/29/2024

A February 22, 2024, FCC draft report and order says the agency should consider low latency a critical requirement of broadband service. The draft states "latency can be critical because it affects a consumer’s ability to use real-time applications, including voice over Internet Protocol (VoIP), video calling, distance learning applications, and online gaming." The FCC notes that there are "many different standards for latency" and references Dave Taht's bufferbloat comments as a "wide-ranging discussion of various potential measures of latency." 

The variety of measures and sources of delay make it difficult to define the broadband criteria as well as an ISP marketing pitch

The report also mentions "consistency of service," which is more readily measured, as a broadband requirement. 

For more on the draft report, click here.


Wednesday, November 01, 2023

Space-ground optical (laser) communication

 

Inter-satellite laser links are in use now, but the technology for optical links to the ground is still being developed and tested. 


Optical frequency laser communication links have many advantages over radio frequency (RF) links:
  • Optical transmission is much faster than RF communication.
  • Optical terminals are smaller and cheaper than RF terminals and use less power.
  •  It's harder to intercept or jam optical signals so they are more secure and, conversely, better for clandestine use.
  • Multiple optical beams can be transmitted simultaneously to multiply speed.
  • Optical transmission is license-free. (There isn't enough RF spectrum to accommodate all of the currently proposed satellites).
SpaceX is equipping its new satellites with inter-satellite laser links (ISLLs). They now have over 8,000 optical terminals in orbit (3 per satellite) and they communicate at up to 100 Gbps. The other low-Earth orbit Internet service providers will follow SpaceX's lead.

Optical communication works well between satellites in the vacuum of space, but optical signals are weakened and distorted by clouds, rain, water vapor, dust, heat gradients, pollen, etc. in the atmosphere so today SpaceX and others use RF frequencies for communication between space and the ground.

Optical space-ground projects

Given the long list of optical advantages, many organizations are working on technology to adjust for atmospheric interference and use optical communication between space and the ground. The following are a few examples of optical communication research and development by NASA, universities, the military, private industry, and the Chinese. 

Ten years ago, NASA demonstrated optical communication between a satellite orbiting the moon and Earth, and they are updating that now. They have a data relay satellite in geosynchronous orbit for relaying data from other satellites to the ground and they are working on transmission from deep space beyond the Moon so we will be able to see video from Mars when we land there. They have also transmitted data between a cubesat with a 2.3 kg payload and the ground at a rate of 200 Gbps. 

ETH Zurich test site
Researchers at ETH Zürich have transmitted data from a mountaintop to their lab 53 kilometers away at up to 0.94 Tbit/s/channel. (Note that the top of the stratosphere is only 50 km from the ground). They adjust for atmospheric variance using sophisticated algorithms and terminals with adaptive optics that can correct the wave phase 1,500 times per second. Their technology can scale up to 40 channels and they are working on more efficient modulation schemes.

The ETH Zürich transmission was from a fixed point on top of a mountain to their lab – how about from a moving satellite? LEO satellites move across the sky at an angular tracking rate of ∼1 deg/s and researchers at the University of Western Australia have demonstrated that they can maintain contact with a drone moving back and forth at that rate. 

In Ukraine, SpaceX Starlink has demonstrated both the military value of satellite Internet and the drawback of being dependent on a private company.

SDA Tracking Layer constellation (source)
The Space Development Agency (SDA) of the Space Force is developing two constellations, a Tracking Layer constellation for warning of, tracking, and targeting advanced missile threats and a Transport Layer constellation providing connectivity to the full range of warfighter platforms. 

There will be ISLLs, using SDA standard optical communication terminals, within and between the early constellations. The early satellites will use RF links to the ground, but optical links are planned.

Tracking plus transport to intercept missiles with planned
optical links to the ground (source)
Space Force policy is to scale up its use of commercial capabilities and Mynaric has been awarded Tracking and Transport Layer contracts for their CONDOR Mk3 optical terminal. CACI International and Tesat terminals have also been certified and will be used -- standards enable competition.

Mynaric has also been selected to participate in a demonstration of links between various space-based optical terminals and an optical ground station they will design. 

The SDA is also working with Aalyria, a startup with two products, Tightbeam and Spacetime, that are based on intellectual property acquired from Google. 

Tightbeam is an optical communication technology that sounds similar to that of ETH Zürich. Using adaptive mirrors and proprietary algorithms, they have transferred data to and from a local mountain at 400 Gbps per channel, (They can use four channels simultaneously). They recently signed a maritime contract for connectivity "starting at" 100 Gbps. 

Tightbeam is only available through Spacetime, an extremely ambitious network operating system for controlling fixed and mobile assets and the links between them on Earth, in the air, and in space. Spacetime runs a simulation of the network and if an upcoming problem is predicted -- for example a weather event or an airplane banking -- Spacetime will reconfigure the network to route around it in 200 ms

(Spacetime is open source with open APIs and Spacetime networks can “federate,” accessing each other’s assets to create a “network of networks.” Sound familiar? APIs are open and they hope to establish standards -- reminiscent of Ethernet vs early proprietary LAN technology. I recommend watching this Spcetime presentation).

Intelsat has provided geostationary satellite communication since the 1960s and is also working with Aalyria on multi-orbit service and space-to-ground optical communication. (They are also considering a medium Earth orbit constellation -- could federating with SES's mPower constellation be an alternative to creating their own)?

I searched for and found two Chinese optical space-ground experiments, one by Beidou in 2021 and a recent test by The Chinese Academy of Sciences with a 10 Gbps transmission rate. I checked with Blaine Curcio, an expert on Chinese space, and he does not know of other tests.

Ground infrastructure

If projects like the above succeed in developing cost-effective space-ground optical communication technology, we will need significant investment in well-designed ground infrastructure. Optical antennas can be added to existing RF ground stations or new optical ground stations can be built.

World cloud cover map (source)
Augmenting existing ground stations makes sense if they are in suitable locations because they already have real estate, power, and Internet connectivity. For example, SpaceX has 75 gateways in North America, several of which are in arid regions of northern Mexico and the US southwest.

New ground stations with optical gateways will also be needed. They should be in relatively cloud-free places and, if possible, near centers of demand and locations with access to high-speed terrestrial Internet connections and power. The current locations of astronomical observatories might be considered.

African gateways
One of the United Nations Sustainable Development Goals is "to build resilient infrastructure, promote inclusive and sustainable industrialization and foster innovation" -- to reduce the digital divide. The needs of underserved areas should also be considered in locating ground stations. Today, SpaceX has only two RF ground stations in Africa, and there are arid regions in the north and south that might be suitable locations for optical ground stations.

ISLL path between arid areas in Mexico and Africa (source).
Such opportunities will increase as ISLLs proliferate.
Even if locations are carefully selected, routing around unfavorable weather or other atmospheric problems will occur at times. That will be facilitated by the proliferation of ISLLs. Furthermore, the addition of ISLLs to sharply inclined orbits will facilitate routing around winter in the northern and southern hemispheres.

Addendum

This post is based on a presentation at a recent UN Internet Governance Forum panel but it has been significantly revised and extended. You can get a copy of the revised PowerPoint presentation here.

The presentation includes a Frequency terminology cheat sheet.

For an excellent tutorial on the properties of laser light, click here.

Thanks to Brian Barrit of Aalyria and Shane Walsh of The University of Western Australia for their input.

Update 11/3/2023

NASA's ILLUMA-T optical terminal will be delivered to the International Space Station in a SpaceX Cargo Dragon launch scheduled for no earlier than November 5. Once installed on the exterior of the space station it will enable two-way communication through the data relay satellite mentioned above to and from optical ground stations in Hawaii and California at 1.2 Gbps.

Monday, September 18, 2023

Will Telesat survive?

In 2017, Telesat, an established Canadian geostationary satellite operator, announced a planned low-Earth orbit Internet service constellation. The plan called for 117 satellites with inter-satellite laser links in a mix of inclined and polar orbits, enabling global coverage. They planned to ignore the consumer market and designed for the enterprise, government, mobile backhaul, mobility, and rural community markets. The mass of their production satellites would be roughly four times that of SpaceX's first-generation satellites. Telesat orbited their first test satellite a month before SpaceX. 

Today, there are over 5,000 Starlink satellites in orbit and Telesat plans to begin launching production satellites in 2026 and won't begin service until 2027. How did Telesat fall so far behind?

Unlike SpaceX, which is fully integrated, Telesat sought vendors for satellite manufacturing, antennas, and launch service, and spent time and money on collaborative design with potential contractors and soliciting and evaluating bids. In February 2021, Telesat selected Thales Alenia Space as the prime contractor for an initial constellation of 298 satellites. (Musk's integrated approach to manufacturing cars or rockets is reminiscent of Henry Ford)

In 2022, Covid, supply chain shortages, inflation, and financing difficulties, led Telesat to cut the constellation to 198 satellites. The schedule slippage raised concerns over Telesat's spectrum licenses (they will have to apply for an extension), and they needed additional capital while their geostationary satellite revenue was declining slightly. Telesat stock, which was offered at $40 per share was trading for a fourth of its initial price and bonds were trading well below par. 

The stock price doubled with the Hail Mary announcements. 
Telesat recently announced two decisions that Tim Farrar, President of TMF Associates, characterizes as a Hail Mary play. 

Last month, MDA, which had been selected to provide Telesat antennas last year, replaced Thales Alenia Space as the prime contractor and manufacturer of the satellites. Telesat expects the new satellites to be about 750 kg -- around the mass of SpaceX's second-generation "Mini" satellites. With these changes, Telesat has funding for the first 156 satellites with the remainder to be paid for from revenue once 156 are in service. 

This month, Telesat announced that they had contracted with SpaceX for fourteen launches starting in 2026. (Note that SpaceX is also launching satellites for competitor OneWeb and Amazon is being sued by shareholders for not considering SpaceX to launch their competing Project Kuiper satellites).

Telesat has lost valuable time. The delay has given SpaceX time to sign over 1.5 million customers, enter the non-consumer markets Telesat is focusing on, and begin launching their second-generation satellites equipped with inter-satellite laser links in both polar and inclined orbits. OneWeb will also be in service and competing with Telesat by 2026. Telesat also says that, despite having geostationary satellites, they will not provide multi-orbit service, while others will

I was on a podcast panel a while back with a Canadian colleague, and he believes the Canadian government will keep Telesat going if needed. I hope Telesat completes its Hail Mary pass -- we need all the competition we can get and the government and military need alternatives to Starlink.

Update 10/19/2023

Telesat explains the benefits of its Lightspeed constellation over rival LEO broadband constellations.

  • Lightspeed satellites orbit at higher altitudes than the others therefore each satellite can serve a larger area with fewer gateways, handoffs are less frequent, and collision risk is lower.
  • Each satellite has four inter-satellite antennas, creating a fixed mesh in space (SpaceX satellites have three).
  • There are only 198 satellites, but Lightspeed antennas can provide nearly 300,000 beams and aggregate up to 15 Gbps within a hot spot like an airport.
  • They will offer service-level agreements including committed information rates.
  • They will work with existing service providers and equipment manufacturers. 

These points are intended to differentiate Telesat from SpaceX, which has a head start in consumer and non-consumer markets, but the stock market is not impressed.

Update 2/26/2024

Avanti Communication, a geostationary satellite operator (GSO) serving government and enterprises in Africa and Europe and Telesat have signed a memorandum of understanding to provide multi-orbit Internet service. Avanti will be marketing Lightspeed service, but the two constellations will operate independently and will not be federating using Aalyria Spacetime. 

GSOs are diversifying by launching non-GSO constellations (SES and Telesat) or buying a low--Earth orbit (LEO) company (Eutelsat-OneWeb). Avanti had decided to market Lightspeed, giving them a LEO product and giving Telesat a customer. Still, Telesat stock continues to fall.

Thursday, July 20, 2023

Google Bard fails to answer satellite Internet questions

Bard made false, inconsistent statements, overlooked pertinent evidence, and overlooked expertise.

In an earlier post, I asked whether electronically steered antennas (ESAs) would replace parabolic antennas in satellite ground stations. I did some research and concluded that it is likely that they will. Next, I discussed the same question with ChatGPT and, while it made several false statements, it made a relevant point that I had overlooked. The relevant addition was positive, but the errors were troublesome, so I decided to try ChatGPT's competitor Google Bard.

Since the P in Chat GPT stands for "pre-trained," I began by asking Bard if it was more up-to-date than ChatGPT and it replied, "Bard is able to access the internet in real-time, while ChatGPT-4 is limited to a dataset that only goes up to late 2021." It added that it had last been updated on June 7, 2023.

That was encouraging since antenna technology has improved since late 2021 and at least two new products incorporating ESAs have been announced since that time, BlueHalo's mobile ground station and Thinkom's gateway array.

But my optimism faded when I began by asking Bard to "list the advantages of electronically steered antennas over parabolic antennas in ground stations for LEO Internet service constellations" and then asked it to "list the advantages of parabolic antennas over electronically steered antennas." The results of the two queries are shown here.

The replies are less complete and specific than those given by ChatGPT, and they are inconsistent -- Bard credits both ESAs and parabolic antennas as being cheaper. Furthermore, there was no indication that Bard was aware of the BlueHalo or Thinkom products, so I asked three leading questions:

  • Have there been recent satellite ground station innovations?
  • Any new hardware?
  • Any innovative new products?
It did not mention either BlueHalo or Thinkom, but after the third try it suggested that I try a Google search for "Innovative new products for satellite ground station" and that returned many links, including one to the Thinkom array of antennas.
My next query demonstrated an extreme lack of awareness. I asked if China's Long March 5B rocket would be used to launch their GuoWang satellite constellation. Bard answered that it would and added that it "will be the first time that a country has launched such a large constellation of satellites." 
Evidently, GuoWang is more frequently associated with Chinese propaganda than the SpaceX Starlink constellation in the Bard training set. (Garbage in, garbage out).
Red and blue text indicates errors (source)

A final example illustrates Bard's inability to recognize expertise. I asked Bard "Who is Larry Press", and it answered with the error-ridden professional biography shown here. The red text highlights statements that are false and the blue highlights work done in collaboration with others, not by me alone.

One can understand some of the errors. For example, I got my Ph.D. from UCLA, not USC, but I did teach at USC for a while and that may be more frequently mentioned in the training set. Similarly, I have been on several editorial boards, but I no longer am. I am an expert on my own biography and a simple Google Search would have found a somewhat dated, but accurate short biography on the Web. 
I was also disappointed in the promise that Bard would provide links to references. It only cites references when it "directly quotes at length from a webpage". I only saw one reference during my experimentation, and it was irrelevant.
"Artificial intelligence" programs have succeeded over the years in many specific tasks. As an undergraduate, I had the privilege of taking a course from Herbert Simon and learned about his chess-playing and theorem-proving programs. Arthur Samuel's checker playing program learned to beat him, and expert systems assist doctors in making some diagnoses. I even wrote an interactive program based on concept acquisition models that assisted researchers in multivariate data analysis. (It could be an Internet service today).
Artificial neural nets have also had success in specific tasks from recognizing zip codes on envelopes to playing Go and Chess and helping robots do back flips and they will succeed at other tasks like helping Microsoft Windows users or writing poetry. (I submitted haikus written by Bard and ChatGPT to the AI detector GPTZero and it reported that they were "likely to have been written entirely by a human").
But none of these programs are generally intelligent.
The illusion of intelligence results from the ability to generate responsive grammatical sentences in a conversation, but that is a low bar. Many years ago, I placed Teletype terminals in a public library and people conversed, sometimes for hours, with ELIZA, a simple simulation of a non-directive therapist. You can try ELIZA for yourself here.
The common practice of referring to errors as "hallucinations" furthers the illusion of intelligence.

Update 9/27/2023
I could not recall the verb "scrub" for navigating through a timeline, so I asked ChatGPT and Bard for help. ChatGPT got it;

Update 10/25/2023

A colleague told me that Bard had cited him as the author of a journal article. I searched for it using Google and could not find it so I asked him about it. It turned out that he had never written an article with the cited title and that the journal it cited had gone out of print two years before the cited publication date.


Friday, June 16, 2023

Will Electronically Steered Antennas Replace Parabolic Antennas in Satellite Ground Stations? (ChatGPT-Assisted Version)

In a previous post, I asked whether electronically steered antennas (ESAs) would replace parabolic antennas in satellite ground stations. I read a few articles suggested by others and by Google search, used some common sense, produced a list of advantages of ESAs, and concluded that it was likely they would eventually replace parabolic antennas for many applications. 

Many of the articles I found were written by companies selling products or services and I'm not an antenna expert -- more a curious journalist. ChatGPT has access to the entire Internet, and I wondered if it could have helped me improve what I wrote or convinced me to reach a different conclusion, so I queried it three times

Since I had listed the advantages of ESAs over parabolic antennas, I began by asking ChatGPT to list the advantages of electronically steered antennas over parabolic antennas for satellite ground stations.

In my post, I listed twelve bullet-point advantages. The ChatGPT answer was more verbose, beginning with a restatement of the question and listing and elaborating on seven advantages. The elaborated replies included all but one of my bullet points, spectral efficiency, and it included an advantage that I had overlooked, interference mitigation, and explained why that was so.

In spite of these glitches, ChatGPT reached nearly the same conclusion as I had, saying "It's worth noting that while electronically steered antennas offer numerous advantages, they also have some limitations, such as higher cost and complexity compared to parabolic antennas. However, ongoing advancements in technology are continuously addressing these challenges, making electronically steered antennas increasingly attractive for satellite ground station applications."

Since it had failed to mention spectral efficiency, I asked, "What about spectral efficiency?" and it replied with an apology for missing the point and went on to list and elaborate on five ways in which ESA made more efficient use of bandwidth.

You can see the dialog here

Since ChatGPT generally agreed with me, I asked it to play devil's advocate and list the advantages of parabolic antennas over electronically steered antennas for satellite ground stations. 

Perhaps I was biased by the experience of having written my earlier post, but I found myself questioning some of its cited advantages. For example, one was "wide coverage" but that would not apply to Thinkom's array of arrays. Another was "simplicity ... making them easier to manufacture, install and maintain." ChatGPT was contradicting its previous response in which it said ESAs were more reliable and required because parabolic antennas "have mechanical parts that can be subject to wear and require regular maintenance."

I also realized that my question was a bit off -- it focused on the antenna rather than the entire ground station -- so I asked it to "list the advantages of satellite ground stations that use parabolic antennas over satellite ground stations that use electronically steered antennas". One of the advantages it listed was suitability for high-frequency Ka and Ku bands, but both Thinkom and BlueHalo use those bands and Thinkom is targeting even higher frequencies. It also repeated the simplicity advantage.

You can see the dialog here.

Finally, I asked the title question from my previous post -- will electronically steered antennas replace parabolic antennas in satellite ground stations?

It listed points under three headings -- advantages of ESAs, limitations of ESAs, and application-specific considerations and concluded that:

"While there is potential for Electronically Steered Antennas to replace parabolic antennas in certain satellite ground station applications, it is not a straightforward or immediate transition. The adoption of ESAs will depend on factors such as technology advancements, cost reductions, performance improvements, and specific application requirements."

It reversed itself once more by listing improved reliability as an advantage of ESAs, but the conclusion was similar to the conclusion in my original post. It also revealed a shortcoming in my formulation of the question. Since I have been writing a lot about Internet service constellations, I overlooked other applications in my previous post and did not specify that I was interested in broadband Internet service. (A human editor would have known that I was writing for an Internet-related publication, would have been aware of my previous writing, and assumed I was focused on Internet applications).

It pointed out that parabolic antennas had an advantage for deep space communication, so I asked a more fully qualified question -- Will electronically steered antennas replace parabolic antennas in ground stations for LEO, MEO, and GEO Internet service constellations?

The reply seemed vaguer and non-committal this time, but was also similar to mine:

"Considering these factors, it is likely that a combination of ESAs and parabolic antennas will be employed in future ground stations for LEO, MEO, and GEO internet service constellations. The specific configuration and utilization of each technology will depend on various factors, including cost, performance requirements, deployment scenarios, and network architectures." 

You can see the dialog here.

Conclusion

So, what is the role of ChatGPT in this sort of journalism? It served me as an editor or referee reviewing what I had written. I would have made two changes after getting feedback from ChatGPT -- I would have mentioned that I was focused on LEO, MEO, and GEO Internet service applications, and I would have included interference mitigation as one of the advantages of ESAs.

ChatGPT made a couple of misstatements and did not "know" about the possibility of an array of antennas like that of Thinkom with its wide elevation angle range. To its credit, I liked the way it apologized for overlooking spectral efficiency as an advantage for ESAs and it was an indefatigable and patient interviewee. It was also free (for the time being).

I consulted ChatGPT after drafting my article, but I could have used it as a research tool before writing. Had I done so, I might have been misled by some of its mistakes and I wouldn't have discovered Thinkom's innovation. Furthermore, its prose was not clear and concise -- I don't think it could pass a Turing Test on writing style. In this case, an old-fashioned Internet search engine was a far better pre-writing tool.

ChatGPT would be improved if it gave links to the sources of its assertions. Like autocomplete, it generates sentences by repeatedly appending the most probable next word in text documents found on the Internet, so the final string is novel. (That's not the way I generate sentences -- they follow from an idea). Might a list of the documents providing the most words be useful? (Google Bard adds citations, but I've not tried it out yet).

While ChatGPT helped me, I don't believe a ground station expert would have learned anything new by interacting with it and a beginner like a student writing a term paper would have been misled. The curious journalist was the sweet spot in this case, but this is version 3.5 of ChatGPT, which had a data-cutoff date before BlueHalo and Thinkom announced the products I've mentioned. I'll revisit it when I get access to ChatGPT4.

PS -- Let me know if you read the dialogs I linked to and notice something I missed,