Wednesday, May 23, 2018

Telesat begins testing low-Earth orbit satellite Internet service

SpaceX and OneWeb get a lot of publicity and have ambitious plans, but Telesat is the first low-Earth orbit (LEO) satellite Internet service provider to begin testing with potential resellers.

Last January, Telesat launched a demonstration satellite and it is now ready for testing. Maritime connectivity provider OmniAccess and Australian ISP Optus had committed to testing the system previously and this week they were joined by in-flight entertainment company Global Eagle Entertainment.

Global Eagle CEO Josh Marks said he was persuaded to collaborate with Telesat by their planned coverage over oceans, polar regions and high-latitude routes and their "open architecture" business model. In addition to testing, they "will collaborate with Telesat on both the technology and commercial model for their new LEO platform.”

OneWeb and several airlines have formed the Seamless Air Alliance, which is developing standards for in-flight Internet connectivity through LEO satellites. I wonder whether Telesat and Global Eagle will join the alliance or go their own way.

Click here for a survey and updated progress report on Telesat and other potential LEO-satellite based Internet service providers.

Monday, May 21, 2018

Elon Musk tells what to expect from the Block 5 Falcon 9 rocket

The ability to launch 30 Falcon 9s per year at a cost of $5-6 million per launch, would be a big plus for SpaceX's Starlink Internet service.

On May 11, SpaceX launched a Bangladeshi satellite using their Falcon 9, Block 5 rocket. This was the first production flight for the Block 5. The day before the launch, Elon Musk participated in a call with reporters and the following are some of the points he made. (You can read more analysis and read a full transcript of the call here)

SpaceX accounted for over half of US launches in
2017 and expects to double their launch rate.
In 2017, SpaceX had 18 successful launches and Musk stated that they were on track to double their launch rate this year, implying a rate of 3 launches per month. He said that "if things go well, which is a caveat, then SpaceX will launch more rockets than any other country in 2018."

There will not be a Block 6. Musk said that after 8 years of upgrades, the Block 5 will be the last major version of the Falcon 9 before their next rocket, the BFR.

Musk expects the Block 5 "to be a mainstay of SpaceX business," and there will be 300 or more Block 5 flights before it is retired in favor of the BFR.

The Block 5 is designed for rapid-turnaround reusability. It is "designed to do 10 or more flights with no refurbishment between each flight — or at least not scheduled refurbishment between each flight. The only thing that needs to change is you reload propellant and fly again." He also said that "the Block 5 boosters are capable of on the order of at least 100 flights before being retired."

Musk has set a goal of demonstrating "two orbital launches of the same Block 5 vehicle within 24 hours, no later than next year."

The Block 5 was designed "to be the most reliable rocket ever built." They have exceeded all of NASA's human-rating requirements and have met "all of the Air Force requirements for extreme reliability."

Reliable reusability will cut cost dramatically. Musk broke down launch cost as follows: booster about 60 percent, upper stage 20 percent, fairing 10% and the launch cost 10%. If they are able to reuse all three rocket elements, they would be able to "reduce the cost for launch by an order of magnitude ... to $5-6 million per launch." Musk pointed out that getting to this point had taken "16 years of extreme effort" (and a lot of learning from failures).

The ability to launch 30 Falcon 9s per year at a cost of $5-6 million per launch, would be a big plus for SpaceX's Starlink Internet service.

Friday, March 30, 2018

FCC approves SpaceX Starlink's Internet-service constellation -- now there are four

On February 18, 2018, FCC Chairman Ajit Pai endorsed the SpaceX application for a constellation of low-Earth orbit (LEO) Internet service satellites and on March 29, the FCC approved their application to "construct, deploy, and operate a proposed non-geostationary orbit (NGSO) satellite system comprising 4,425 satellites for the provision of fixed-satellite service (FSS) around the world."

That makes SpaceX the fourth company with permission to operate an LEO Internet service constellation in the U. S.

The first was OneWeb on June 22, 2017. OneWeb received permission to deploy 720 LEO Internet-service satellites, subject to an important constraint that they "need to accommodate in-line interference avoidance and spectrum sharing with other NGSOs in the future." That cleared the way for spectrum sharing among all operators.

The applications of Telesat and Space Norway were both approved on November 2, 2017. Telesat was granted permission "to access the U.S. market to provide FSS using a proposed constellation of 117 NGSO satellites" and Space Norway was granted permission to "to access the U.S. market to provide FSS using a proposed constellation of two NGSO satellites." (Space Norway is planning coverage in the area north of 65 degrees N latitude, which includes northern Alaska).

Wednesday, March 14, 2018

O3b satellite Internet -- today and tomorrow

I have written a lot about the potential of low-Earth orbit (LEO) satellites for Internet service, but have not said much about medium-Earth orbit (MEO) satellites -- until now.

O3b (other three billion) is an MEO-satellite Internet service provider. Greg Wyler founded the company and it was subsequently acquired by SES, a major geostationary-orbit (GSO) satellite company. (Wyler moved on to found future LEO Internet service provider OneWeb).

O3b's MEO satellites orbit at an altitude of around 8,000 kilometers, complementing the SES GSO constellation, which orbits at around 36,000 km. Because of their altitude, the GSO satellites have large footprints and are good for video broadcast and other asynchronous applications, but their latency can cause noticeable delays in applications like telephony or browsing modern, complex Web sites which may have many elements -- images, text, video, programs, etc. -- each adding transmission and error-checking overhead.

The International Telecommunication Union says that if one-way latency is less than 150 milliseconds, users of most applications, both speech and non-speech, will experience essentially transparent interactivity. I've never used an O3b link, but they promise a round-trip latency of under 150 milliseconds, so I assume they work well for voice calls, where GSO satellites would introduce a perceptible delay. However, an MEO link might be noticeably slower than a LEO link while browsing modern Web sites.

O3b launched four satellites last week and they plan to launch four more early next year. That will bring their constellation to 20 satellites and enable them to continue expanding their current business of serving relatively large customers like mobile phone companies, government organizations, and cruise ship lines. For example, they serve Digicell which has over 40,000 LTE accounts in Papua New Guinea.

There is a growing market for O3b's current service, but their next-generation satellite-communication system, called mPOWER, will be much improved and will compete in some applications with terrestrial fiber and, in the future, with LEO constellations. The initial mPOWER constellation of seven satellites was designed and will be manufactured by Boeing. While today's O3b satellites have 10 steerable edge-terminal beams, the mPOWER satellites will have over 4,000 steerable beams that can be switched under program control giving the constellation over 30,000 dynamically reconfigurable beams and over 10 Tbps capacity. The highly-focused beams will address single terminals, not wasting power on locations with no users. The constellation will be launched in 2021.

Seven satellites, each with over 4,000 steerable, fully-shapeable beams

O3b has also contracted with three customer-edge terminal manufacturers ALCAN, Isotropic Systems and Viasat. I don't know the prices or capabilities of these terminals, but it is probably safe to say they will utilize different, novel antenna technologies and have different prices and characteristics for different applications. I am guessing that ALCAN is working on a low-cost, flat-antenna terminal using their liquid crystal technology; Isotropic is working on antennas based on their optical beamformer modules and Viasat is working on a terminal with a more expensive phased-array antenna that can communicate with both SES GEO and mPOWER MSO satellites.

In the press conference announcing mPOWER, SES Networks CEO emphasized that these were smart terminals -- computers that happened to have antennas. They will be continuously connected and monitored and able to download software updates and new applications. They will be parts of an integrated system comprised of edge terminals, terrestrial POPs and gateways, SES GSO satellites, existing O3b and future mPOWER satellites and terrestrial fiber and wireless networks. The network will be dynamically configured under program control as a function of applications and cost.

Applications for flat-panel edge terminals (source)

The first seven mPOWER satellites will be in equatorial orbit and cover nearly 400 million square kilometers between + and - 50 degrees latitude. Once mPOWER is up and running, SES plans to retire and not replace two of their GSO satellites and they will add to the initial seven-satellite constellation when the capacity is needed. They will still market to customers like governments, ISPs, mobile phone companies and ship and airlines, but will be able to serve many more and smaller organizations.

As I mentioned at the start of this post, I have been covering LEO Internet-service satellite projects for some time, and the two leading contenders are SpaceX and OneWeb. SpaceX launched their first two test satellites last month and they plan to start launching operational satellites in 2019. They will begin offering commercial service in the 2020-21 time frame and complete their first global constellation by 2024. OneWeb plans to begin offering service in Alaska in 2019 and to cover all of Alaska by the end of 2020. By 2025 they expect to have 1 billion subscribers.

O3b mPOWER will be up and running before SpaceX and OneWeb complete their global constellation -- they may even have launched their second batch of satellites.

Sunday, January 14, 2018

Courts can combat gerrymandering with gerrymandering tools

Gerrymandering -- defining voting districts to favor one party or candidate -- has been with us for years, but it was difficult to do and imprecise. Mapping software using Internet voter data have made it precise and easy.

Gerrymandering is in the news because a panel of federal judges ordered North Carolina to redraw its gerrymandered congressional map.

The panel struck down North Carolina’s congressional map, saying it was unconstitutional because it violates the 14th Amendment guarantee of equal protection. Judge James A. Wynn Jr., in a biting 191-page opinion, said that Republicans in the North Carolina legislature had been “motivated by invidious partisan intent” as they carried out their obligation in 2016 to divide the state into 13 congressional districts, 10 of which are held by Republicans.

The ruling will be appealed directly to the Supreme Court, which is also hearing Wisconsin and Maryland gerrymandering cases. The Wisconsin and South Carolina cases are both based on the 14th amendment and are pro-Democratic while the Maryland case challenges the redrawing of a single district, is based on the 1st Amendment and is pro-Republican.

Gerrymandering is not new -- Patrick Henry tried to defeat James Madison in 1788 by drawing an anti-federalist district. He failed because he did not have good data and computers, but today's politicians have geographic information system software and the data they need to automate efficient, precise gerrymandering. (The term "gerrymandering" was coined in 1812 when Massachusetts governor Elbridge Gerry reluctantly approved a map in which one district resembled a salamander).

The Republican party has used Internet-enabled gerrymandering to gain a congressional advantage. The Democratic party might be tempted to fight fire with fire, but that would be slow and undemocratic.

The North Carolina judicial panel has a better solution. They gave the legislature until January 24 to present a “remedial plan” and the court will institute its own map if it finds the new district lines unsatisfactory. If that happens, the court can use use the same sorts of tools and data that have been used to produce gerrymandered districts. Instead of using the technology to optimize in favor of either party, they will seek maps that equalize district populations, minimize geographic perimeters, respect natural boundaries like rivers, maximize racial diversity, etc. In general, courts are more likely to be non-partisan than legislatures.

Update 1/19/2018

Pennsylvania county results, 2016
The U. S. Supreme Court granted a stay in the court order requiring North Carolina lawmakers to produce a revised congressional voting map within two weeks. This temporary delay probably means the current map will be used in the 2018 election.

In a related case, the Pennsylvania state supreme court is currently hearing a gerrymandering case which could result in the redrawing of their district map in time for the 2018 election.

Republicans won 13 of Pennsylvania's 18 seats in the U.S. House of Representatives in 2016 in spite of the fact that Donald Trump defeated Hillary Clinton by only 44,292 (.75%) votes.

Update 1/30/2018

The Pennsylvania Supreme Court has decided that the Republican-drawn district map violates the State Constitution and ordered that it be redrawn by February 9th.

Republicans hold 72% of Pennsylvania's 18 seats in Congress and only carried the state by .75% in the last election. That imbalance raised a red flag, but there are no hard and fast rules for determining whether oddly-shaped districts or disproportional representation are due to intentional gerrymandering or other factors like compliance with the 1965 Voting Rights Act, which prohibits racial discrimination, or people choosing to live in homogeneous neighborhoods.

A "legitimate" district
In a sense, the judges had to decide the intent of the Republicans in drawing their district map and the following exchange during the hearing may have led them to rule that the intent was to gerrymander:
Justice Max Baer: “if you took the Democratic areas of Pittsburgh and Philadelphia and connected them via the Pennsylvania Turnpike, that’s okay?”

Jason Torchinsky, a lawyer representing the Pennsylvania Republican party: "yes.“
That would be an extreme example of "packing" -- putting all of Pennsylvania's urban Democrats into a single, overwhelmingly Democratic district. A Democrat would win the packed district in a landslide, thereby "wasting" many Democratic votes. Our Electoral College voting system also disenfranchises voters in predominantly Republican or Democratic states, thus giving inordinate power to a handful of competitive "swing" states in national elections.
Illinois District four
While the hypothetical Pennsylvania example is blatant, an unusually shaped district could also be the result of compliance with the Voting Rights Act. For example, Illinois District four has two areas connected by an uninhabited stretch of land along Interstate 294, creating a Hispanic-majority district.

For a funny and informative discussion of the difficulty of drawing fair district maps and another blatant example of partisan bias, watch the following video, but be forewarned that it includes some adult content.

Update 2/12/2018

Last Friday, Pennsylvania Republicans submitted their proposed district map to the Pennsylvania Supreme Court. The governor says he will review the map and decide whether or not to recommend that the Court accept it by February 15th. If he rejects the map, the Court will impose its own map on February 19, in time for primary elections.

In rejecting the current map, the Court noted that in the three elections held after the last redistricting, Democrats won the same five seats and Republicans won the remaining 13 seats every year, in spite of the fact that the Democrats won between 46 and 51 percent of the statewide popular vote in each election. Furthermore, in 2016, Democrats won their House seats with an average of 75 percent of the vote, while the Republicans' victory margin averaged only 62 percent. The Court concluded that the Democrats had been packed into five districts, "wasting 25 percent of their votes," and the Republicans were spread out among the remaining 13 districts.

This map shows the packing of the seventh district:


The Court found that the evolution of the district from a compact, contiguous region at the time of the 83rd Congress (1953-5) to today's map indicated illegal gerrymandering:


The stakes are high. If redistricted, Pennsylvania might end up with nine Republican and nine Democratic representatives and that could make the difference between Democratic or Republican control of the US House of Representatives.

Update 3/24/2018

Map redrawn by the Supreme Court
The Pennsylvania Supreme Court rejected the redrawn map submitted by the Legislature, substituting the map shown here which has:
  • 10 Trump 2016 districts
  • 8 Clinton 2016 districts
  • 1908 district border miles
  • 13 county splits
As compared to the old map which had:
  • 12 Trump 2016 districts
  • 6 Clinton 2016 districts
  • 3047 district border miles
  • 28 county splits
A dozen Republican members of the Pennsylvania House of Representatives have filed impeachment resolutions against the five Supreme Court judges who are Democrats, arguing that they had violated the State Constitution.

The Chief Justice of the Supreme Court, a Republican, released a statement calling the impeachment resolution "an attack upon an independent judiciary."

Two New York Times staff members drew the following map to demonstrate the ease with which partisan redistricting is achieved using mapping software with data available on the Internet. They did it using Dave's Redistricting Service -- you can try it for yourself.

A blatantly partisan, hypothetical district map drawn using an Internet service

It is easier to spot gerrymandering and state the problem than it is to come up with an equitable, constitutional solution, but we should do our best to keep partisan politics out of the process and legislators are nearly all members of a party. Having courts draw up maps, as was done in Pennsylvania, or allowing states to establish independent redistricting commissions like the one in California seem like imperfect steps in the right direction.

Thursday, December 14, 2017

Eighteen posts on low-Earth orbit satellite Internet service

The internet is unavailable to and/or unaffordable by about 50% of the world population. The situation is worse in, but not confined to, developing nations where the service is typically sub-standard when it is available.

Percent of homes with Internet connectivity (source)

Geostationary satellite connectivity is available globally, but it is slow and expensive because the satellites are high above the Earth. Low-Earth orbit (LEO) satellites can deliver speeds comparable to terrestrial links, but constellations of many satellites would be needed to serve the entire planet.

The first project to attempt a LEO constellation failed in the 1990s, but rocket, electronic and communication technologies have made great strides since that time. Today, five LEO satellite-Internet projects that hope to provide global, affordable, high-speed Internet are underway. If they succeed, we will see early LEO connectivity in some places (like Alaska) in 2020 and by the middle of the next decade homes, schools, libraries, businesses, ships at sea, Internet-connected devices, etc. will be online. While mobile connectivity is growing rapidly in developing nations, high-speed fixed connectivity would enable the use of personal computers, a qualitative improvement for content creators.

I've watched LEO satellite connectivity since the 1990s, but technological progress has led to renewed interest and investment in recent years, leading me to follow the developments in my class and on our class blog.

The following are annotated links to 18 blog posts written since June 2014. The listed dates show when the post was first published, but each has been updated several times since publication. (For example, I updated the posts on Boeing and OneWeb's projects this week). The posts are illustrated with around 80 images and include 15 videos of important talks and events. When appropriate the posts link to other posts within the collection, creating a document that can be read sequentially or as a hypertext. Regardless, I would suggest starting with the most recent post. There are also hundreds of links to external sources.

  • An example of effective government support for new communication technology. (November 2017)
  • From the Morse telegraph through the Internet, government and industry have collaborated on the development and deployment of communication technology. A recent hearing by the Senate Commerce, Science, and Transportation Committee continued that cooperation by seeking suggestions for helpful legislation from four satellite industry representatives. The constructive, non-partisan tone of that hearing stands out in the current era of polarized, dysfunctional government. For more on the testimony of witnesses from SpaceX and OneWeb, see: SpaceX and OneWeb.

  • Telesat -- a fifth satellite Internet competitor. (November 2017)
  • Telesat hopes to achieve global coverage rapidly and at low cost by deploying a small, hybrid constellation with both polar and inclined-orbit satellites connecting to the terrestrial Internet via ground stations they already own in the far north. Like OneWeb, they are working with outside vendors for launch services, satellite and antenna design and manufacture. This and the size of their constellation will keep initial capital costs relatively low.

  • Will low-Earth orbit satellite Internet service providers succeed? (October 2017)
  • There is good news and uncertainty/bad news. Teledesic failed as a satellite ISP, but since that time we have seen vastly improved technology and changes in the terrestrial Internet industry, market and executive experience. These changes are generally positive, but the new satellite companies also face unique technical, political and business roadblocks and unknowns. The future is uncertain, but increased Internet service competition would benefit us all.

  • The BFR and its role in SpaceX's satellite Internet service. (October 2017)
  • This post is based primarily on a talk (with an excellent slide deck) that Elon Musk gave on their forthcoming Big Falcon Rocket (BFR). He outlined its specifications, for example, ten times the payload capacity of the current Falcon 9, and the ways it will be used for inserting satellites in orbit and establishing a base on Mars. The Falcon 9 will be used to launch SpaceX's first two prototype Internet-service satellites early next year, but in 2019, when they begin launching operational satellites, their next rocket, the Falcon Heavy, will be available. The BFR will be available before their first Internet constellation is complete in 2024. Musk suggested that, in addition to launching terrestrial satellites and travel to Mars, the BFR would be used to retrieve spent satellites and second stages and for long-distant terrestrial travel.

  • Non-terrestrial spectrum sharing. (October 2017)
  • Satellites and terrestrial Internet service providers rely on radio communication and must avoid interference when transmitting near each other at a given frequency. Historically, this has been achieved by granting exclusive licenses to use specific frequency bands, but this is not practical or efficient with thousands of satellites at different altitudes and in different orbits. Fortunately, modern communication technology opens the possibility of dynamically sharing frequencies among many providers -- terrestrial and satellite -- but cooperation and standards are needed. The satellite Intenet companies embrace frequency sharing and are willing to cooperate.

  • Can constellations of Internet-routing satellites compete with long-distance terrestrial cables? (September 2017)
  • Packets on routes between distant points are relayed through multiple routers and each inter-router "hop" takes time. Generally speaking, satellites, which can see far over the horizon and make straight-line connections to other satellites, require fewer router hops than fiber links between distant points. Furthermore, laser transmission in space is faster than in fiber. For these reasons, Elon Musk has set a goal of having "the majority of long-distance traffic go over this (satellite) network" and Leosat is focusing on high-end fast, point-point links.

  • Boeing's satellite Internet project. (August 2017)
  • Boeing has been in the satellite business for many years -- they were the prime contractor for Teledesic's failed attempt at LEO satellite Internet in the late 1990s. They have applied for a license to launch 1,396 satellites within six years and another 1,560 within 12 years, but have kept a relatively low profile. They recently shortened the timetable on their Mars exploration project so the large Space Launch System rocket they are developing for Mars voyages may become available for launching Internet satellites. They may also be working toward collaboration with OneWeb

  • OneWeb satellite Internet project update. (August 2017)
  • This post outlines OneWeb founder Greg Wyler's background in the terrestrial and satellite Internet service business and presents the current status of their Internet project. Softbank is a leading partner and investor and Wyler gave a talk at the 2017 Softbank World conference in which he discussed their technology, major partnerships, anticipated timeline, goals and the advantages they enjoy. They plan to launch their first satellites in March 2018 and begin offering service in Alaska in 2019. They hope to cover all of Alaska by the end of 2020 and have 1 billion subscribers by 2025. Their prices will vary with regional incomes, so they hope to eliminate the global digital divide by 2027. The post includes a video of Wyler's talk at the Softbank conference and his testimony before the Senate Committee on Commerce, Science and Transportation.

  • SpaceX satellite Internet project status update. (August 2017)
  • Patricia Cooper, SpaceX Vice President, Satellite Government Affairs, testified at a productive hearing by the Senate Committee on Commerce, Science, and Transportation. She outlined SpaceX's plans and gave a rough timetable for two constellations -- one in low-Earth orbit and a larger constellation in very-low Earth orbit, which may position them to compete with terrestrial ISPs in densely populated urban areas and serve the so-called "Internet of things." (Tesla cars and solar roofs would be likely things to connect). Elon Musk has given a talk outlining the SpaceX timetable for establishing early settlements on Mars and giving some details on their future booster rockets, the Falcon Heavy and the Big Falcon Rocket (the BFR). The first Falcon Heavy is expected to be launched in January 2018 (the payload will be a Tesla Roadster). Musk did not say when the BFR will be ready, but it will be before their first Internet-satellite constellation is complete. He also said the BFR might be used to retrieve spent satellites. SpaceX has also trademarked "Starlink" as the name of their satellite Internet service.

  • Cool images and video of the latest, increasingly routine SpaceX soft landing. (May 2017)
  • SpaceX has made landing a 549,054 kg rocket that is 70 feet long and only 3.66 meters in diameter and has reached an altitude of 247 km and fallen at a speed of up to Mach 7.9 within .7 m of the target on a drone barge at sea almost routine. Recovering and reusing boosters and eventually, second-stage rockets and satellites will dramatically reduce cost and downtime. This post documents early failures and eventual success.

  • Two approaches to routers in space -- SpaceX and OneWeb (February, 2017)
  • SpaceX and OneWeb have the same goal, but their organizations are dissimilar. SpaceX is integrated -- building the rockets, satellites and ground stations themselves -- while OneWeb has a number of collaborators and investors, including Bharti Enterprises, Coca-Cola, Intelsat, Hughes, Totalplay Telecommunications, Virgin Galactic and Softbank. OneWeb attempted a merger with Intelsat, which would have given them international offices and access to geostationary satellites, but the merger failed.

  • Satellite Internet update -- Airbus will make satellites for OneWeb. (June 2015)
  • Airbus will make satellites for OneWeb in a joint-venture factory in Florida and Softbank has invested $1 billion. This post features an in-depth interview of Brian Holz, OneWeb’s Director of Space Systems, who speaks about the reasons for producing the satellites in the US and the factors in choosing a factory location, the cost of the satellites ($4-500,000 each), the need to have global participation in a global project, launch services, satellite reliability and plans for eventually deorbiting them, financing and the business case, the search for manufacturers of millions of user terminals and antennas, etc.

  • SpaceX is ready to test Internet service satellites. (June 2015)
  • SpaceX filed an application to launch two identical test satellites to validate the design of their broadband antenna communications platform using three broadband array test ground stations along the western coast of the US. OneWeb seems to be moving faster.

  • Greg Wyler reports OneWeb progress. (March 2015)
  • OneWeb has found major partner/investors and is developing a $250, user-installable ground station that will serve as a WiFi hotspot and a 2G, 3G or LTE cell station. Consider the possibility of a WiFi network with a low-latency, 50 Mbps back-haul link to the Internet in every school or rural clinic in the world.

  • Leosat -- a third satellite Internet company. (March 2015)
  • Leosat will focus on the high-end market, providing low-latency, high-speed, secure connectivity to government and business -- maritime applications, oil and gas exploration and production, telecom back-haul and trunking, large enterprises, etc.

  • Regulation of global satellite Internet service providers. (January 2015)
  • Would global Internet service providers require unique regulation and, if so, what should it be and who has the power to do it?

  • Elon Musk and Greg Wyler's plans for global satellite connectivity. (November 2014)
  • Greg Wyler first tried to bring fiber connectivity to Rwanda, but, when that proved difficult he turned to satellite, founding O3b, a medium-Earth orbit Internet satellite company with the goal of connecting "other three billion." Elon Musk, founder of SpaceX, Tesla and other companies needs no introduction. (I am a Musk fan). Google invested in O3b and Wyler worked there for a while, then contemplated a low-Earth orbit constellation in partnership with Musk, but eventually formed his own company, OneWeb. This post covers the visions and early efforts of Musk at SpaceX and Wyler at OneWeb.

  • Can Google connect the "other three billion" in developing nations and rural areas? (June, 2014)
  • A survey of Google's early work with high-altitude platforms (Project Loon), low-Earth orbit satellites, medium-Earth orbit satellites and geostationary satellites

Wednesday, November 29, 2017

An example of effective government support for new communication technology

Based on their questions and comments during the Senate Commerce, Science, and Transportation Committee hearing on the commercial satellite industry, one could not tell whether a senator was a Democrat or Republican.

The US government has a history of support of telecommunication. On March 3, 1843, the US Senate passed a bill "to test the practicability of establishing a system of electro magnetic telegraphs by the United States." The bill provided $30,000 for Samuel Morse to conduct the test. He built a telegraph link between Washington and Baltimore and the rest is history.


US government R&D, procurement, regulation, and expertise also played an important role in the development of the Internet -- see Seeding Networks, the Federal Role. (If you do not have access to the paper, send me a request for a copy). Government collaborated with universities and industry on the development of the Internet up to the time they phased out support, as shown below:

Federal funding prior to the NSFNet phase out
Source: my ACM bibliography (scroll down)

The October Senate Commerce, Science, and Transportation Committee hearing on the commercial satellite industry provides a current example of effective government support of new communication technology.

The hearing focused on broadband access, primarily from low-Earth orbit (LEO) satellites. Witnesses from four companies -- Intelsat, OneWeb, ViaSat and SpaceX -- testified and the tone of the hearing was set by the opening statements of Committee Chairman John Thune and Ranking Member Bill Nelson. Thune began by saying "I believe we are at a critical moment in the development of satellite capability, and I am excited to hear from our panel of distinguished witnesses today." In his opening remarks, Nelson echoed Thune's optimism and among other things stated that he "would like to thank our witnesses for being here today and I look forward to discussing how we can work together to bring about this new Space Age."

The senators were sincere in their desire to serve the American people and they were asking for recommendations as to how they could craft legislation to realize the potential of satellite broadband service. A short introductory statement by each witness was followed by questions and answers. The senator's questions were constructive -- trying to learn from the witnesses, not score political points with their constituents. Based on their questions and comments, one could not tell whether a senator was a Democrat or Republican. They were all constructive.

I was also struck by the degree of overlap in the recommendations given by the four executives, for example:
  • They are all in favor of sharing spectrum among themselves and with terrestrial service providers. They agree that dividing frequency bands among operators is the least desirable and most inefficient way to avoid interference.
  • The four agree that satellite safety and debris mitigation will be critical in an era of large constellations of LEO satellites and that we need to work with International agencies to establish standards. They understand that a disastrous collision would set the entire industry back so they have a common interest in satellite safety.
  • Global standards are needed for debris mitigation, spectrum sharing, etc. and the US, with its history and expertise at NASA and the staffs of agencies like the FCC and NTIA, can and should take the lead in establishing those international standards.
  • The government definition of "broadband Internet" should be technology neutral. Today's geostationary satellite service is slower than terrestrial service, but speeds of coming LEO services will be comparable to terrestrial service.
  • OneWeb is also working on a grappling mechanism for retrieving spent satellites and Greg Wyler said they "hope to open source" the design. No other witnesses mentioned open source, but given Tesla's open source policy, we might expect open source designs from SpaceX as well.
The only explicit disagreement I heard was OneWeb arguing against the Connect America Fund subsidy, but, if it is not limited, I am sure they would like to receive funds. In general, the satellite providers have many common interests and they would like procedures and policies adjusted to allow them to compete on a level playing field with terrestrial ISPs.

Watching this hearing reminded me of the collaboration between Intel, Digital Equipment Corporation and Xerox to create the Ethernet standard. Potential competitors grouped together to define a standard that would enable a large, competitive market as opposed to several small proprietary markets. LEO satellite broadband feels like a startup industry -- reminiscent of the early personal computer days in the US, the ARPAnet and Internet in the days of the Acceptable Use Policy or even the Cuban start-up scene today.

You can see the hearing yourself -- the senators' opening statements, the written testimony of the witnesses and a video of the entire hearing, including questions, answers, and discussion among the senators and witnesses may be found here. SpaceX and OneWeb are both planning large LEO satellite constellations and you will find summaries of their testimonies at the end of these progress reports: SpaceX and OneWeb.

Monday, November 13, 2017

Telesat -- a fifth satellite Internet competitor

Telesat will begin with only 117 satellites while SpaceX and the others plan to launch thousands -- how can they hope to compete? The answer lies in their patent-pending deployment plan.

Polar (green) and inclined (red) orbits
I’ve been following SpaceX, OneWeb, Boeing and Leosat's satellite Internet projects, but have not mentioned Telesat's project. Telesat is a Canadian company that has provided satellite communication service since 1972. (They claim their "predecessors" worked on Telstar, which relayed the first intercontinental transmission, in 1962). Earlier this month, the FCC approved Telesat's petition to provide Internet service in the US using a proposed constellation of 117 low-Earth orbit (LEO) satellites.

Note that Telesat will begin with only 117 satellites while SpaceX and the others plan to launch thousands -- how can they hope to compete? The answer lies in their patent-pending approach to deployment. They plan a polar-orbit constellation of six equally-spaced (30 degrees apart) planes inclined at 99.5 degrees at an altitude of approximately 1,000 kilometers and an inclined-orbit constellation of five equally-spaced (36 degrees apart) planes inclined at 37.4 degrees at an approximate altitude of 1,248 kilometers.

This hybrid polar-inclined constellation will result in global coverage with a minimum elevation angle of approximately 20 degrees using their ground stations in Svalbard Norway and Inuvic Canada. Their analysis shows that 168 polar-orbit satellites would be required to match the global coverage of their 117-satellite hybrid constellation and according to Erwin Hudson, Vice President of Telesat LEO, their investment per Gbps of sellable capacity will be as low, or lower than, any existing or announced satellite system. They also say their hybrid architecture will simplify spectrum-sharing.

The following figure from their patent application illustrates hybrid routing. The first hop in a route to the Internet for a user in a densely populated area like Mexico City (410) would be to a visible inclined-orbit satellite (420). The next hop would be to a satellite in the polar-orbit constellation (430), then to a ground station on the Internet (440).

An inter-constellation route (source)

The up and downlinks will use radio frequencies and the inter-satellite links will use optical transmission. Since the ground stations are in sparsely populated areas and the distances between satellites are low near the poles, capacity will be balanced. This scheme may result in Telesat customers experiencing slightly higher latencies than those of their competitors, but the difference will be negligible for nearly all applications.

They will launch two satellites this year -- one on a Russian Soyuz rocket and the other on an Indian Polar Satellite Launch Vehicle. These will be used in tests and Telesat says a number of their existing geostationary satellite customers are enthusiastic about participating in the tests. They will launch their phase 2 satellites beginning in 2020 and commence commercial service in 2021. They consider 25 satellites per launch vehicle a practical number so they will have global availability before their competitors. Their initial capacity will be relatively low, but they will add satellites as demand grows.

Like OneWeb, Telesat will work with strategic partners for launches and design and production of satellites and antennae. They have not yet selected those partners, but are evaluating candidates and are confident they will be ready in time for their launch dates. Their existing ground stations give them a head start. (OneWeb just contracted with Hughes for ground stations).

Their satellites will work with mechanical and electronically steered antennae and each satellite will have a wide-area coverage mode for broadcast and distributing software updates. Their patent application mentions community broadband and hotspots, large enterprises, ships and planes, software updates and Internet of things, but not homes as initial markets.

Telesat's Canadian patent application goes into detail on all of the above, and I'd be curious to know what exactly would be protected by it. They also consider their global spectrum priority rights from the International Telecommunication Union as an asset, but they will have to agree to spectrum sharing conventions and debris mitigation agreements.

Let me conclude with a suggestion for Telesat and the Cuban government.

OneWeb has committed to providing coverage to the entire state of Alaska by the end of 2020 and Telesat says they will have global coverage by 2021. I follow the state of the Internet in Cuba and think Cuba would be a good starting place for Telesat service. Cuba has the best-educated, Internet-starved population in Latin America and the Caribbean, they have very little domestic Internet infrastructure and much of the infrastructure they do have is obsolete. Cuba is close to being an Internet "green field" and, since it is an island nation, their polar satellite "footprint" would not be densely populated.

Cuba could work with Telesat to leapfrog over several infrastructure generations. If Telesat can deliver on their claims, the barriers would be political and bureaucratic, not technical. Cuba is about to change leadership, and there is some indication that Miguel Díaz-Canel, who many expect to replace Raúl Castro, will favor Internet development.

SpaceX could also provide early Cuban connectivity, but dealing with a US company would be politically problematical and Cuba and Canada have a well established political and economic relationship. Even if Cuba were willing to work with SpaceX, the current US administration would not allow them to do so. Connecting Cuba would be good for Cubans and good publicity of Telesat.

For more on Telesat and their plans for LEO satellite Internet service see their patent application and you can see animations of their proposed hybrid-constellation connectivity here and here.

Update 11/25/2017

LEO-1, Telesat's low-Earth orbit satellite, has been shipped to India for launch. The 168 kg satellite will be used in two-satellite tests of Telesat's forthcoming broadband service. Testing will begin when both test satellites are in operation. LEO-1 will be in polar orbit and I assume the other will be in an inclined orbit in order to test their two-constellation design.

Update 11/29/2017

The Soyuz 2 launch vehicle that was to have placed 19 spacecraft into orbit has failed, destroying one of the two satellites Telesat had planned to use in the first test of their forthcoming broadband Internet service. The other has been shipped to India for launch, but the project will be delayed until the lost satellite can be replaced.

Update 5/19/2018

SpaceX and OneWeb get a lot of publicity and have ambitious plans, but Telesat is the first LEO ISP to begin testing with potential resellers.

Last January, Telesat launched a demonstration satellite and it is now ready for testing. Maritime connectivity provider OmniAccess and Australian ISP Optus had committed to testing the system previously and this week they were joined by in-flight entertainment company Global Eagle Entertainment.

Global Eagle CEO Josh Marks said he was persuaded to collaborate with Telesat by their planned coverage over oceans, polar regions and high-latitude routes and their "open architecture" business model. In addition to testing, they "will collaborate with Telesat on both the technology and commercial model for their new LEO platform.”

OneWeb and several airlines have formed the Seamless Air Alliance, which is developing standards for in-flight Internet connectivity through LEO satellites. I wonder whether Telesat and Global Eagle will join the alliance or go their own way.

Tuesday, October 24, 2017

Will low-Earth orbit satellite Internet service providers succeed?

Teledesic animation: a satellite
constellation that would cover
the planet -- routers in space.
In 1990, Teledesic was formed to deliver satellite-based Internet service. Cellular pioneer Craig McCaw, Microsoft co-founder Bill Gates and Saudi Prince Alwaleed bin Talal were early investors and Boeing was both an investor and the prime contractor. Teledesic hoped to offer global Internet connectivity using a constellation of 840 satellites in low-Earth orbit (LEO) at an altitude of 700 km. (The plan was scaled back to 288 satellites in 1997).

Teledesic failed.

Twenty seven years later three companies SpaceX, OneWeb, Boeing and Leosat are trying to do what Teledesic could not do. Will they succeed?

Good news -- a lot has changed since Teledesic tried and failed.

Launches have gotten cheaper -- SpaceX has made landing a 549,054 kg rocket that is 70 feet long and only 3.66 meters in diameter and has reached an altitude of 247 km and fallen at a speed of up to Mach 7.9 within .7 m of the target on a drone barge at sea almost routine. They have had 18 successful soft landings and their next-generation rocket, the BFR is designed for reusability and will be able to launch many satellites per flight.

Satellites are now cheaper, smaller and lighter. OneWeb and their manufacturing partner Airbus say automation and re-design will enable them to manufacture three satellites per day at a cost of less than $1 million each and launch cost per satellite will be low since they are small and light. In a talk at the opening of the SpaceX satellite engineering office in 2015, CEO and Lead Designer Elon Musk expressed confidence that they would be able to mass produce satellites and also pointed out that the failure of one or a few satellites in a constellation of thousands was relatively unimportant. If a satellite fails, the remaining satellites will route around it and increased tolerance of failure reduces manufacturing cost.

Consumer ground-stations will also be small, cheap and easy to install "pizza boxes." Since the signals will be weak, the antennae will have to be outside, but end users will be able to install them because, unlike today's TV and Internet dishes, they will not have to be aimed precisely at a single, geostationary satellite. Modern phase-shift array antennae will follow a satellite as it moves and switch to another in microseconds or a few milliseconds when it goes out of site. Similarly, satellites will be able to transmit a beam that is fixed on one area on the ground as it moves over it.

Communication technology has improved dramatically. Today's radios are smart -- rapidly changing power, frequency, modulation scheme, etc. under program control. Teledesic belonged to an era of dedicated spectrum allocation, but smart radios are ushering in an era of unlicensed spectrum and spectrum sharing. This is a fundamental shift -- like the introduction of packet switching -- and it will lead to efficient use of spectrum (on Earth and in space).

The terrestrial fiber network has grown exponentially since the time of Teledesic. making Internet gateways attractive targets. Satellites may compete favorably with long undersea and terrestrial cables. Elon Musk says SpaceX satellites will communicate optically among themselves, forming a low-latency, highly interconnected mesh that will carry the majority of our long-distance traffic. Google, a billion dollar SpaceX investor, might be their first long-haul customer.

The market for Internet connectivity is larger today than it was at the time of Teledesic because more people are aware of the Internet and many are trained to use it with mobile devices. Furthermore, applications are varied and powerful today. The Teledesic project was during the Web 1.0 era when the Web consisted of static sites with text and small, compressed images. Today, the U. S. Federal Communication Commission (FCC) defines "broadband" as at least 25 Mbps download speed and some fortunate people have gigabit connections in their homes. The low latency of LEO satellites is well suited to interactive applications that require error checking.

While increasing numbers of people are online with mobile devices, the digital divide has widened. A slow connection using a phone is not the same as a fast connection that can support modern applications on a computer. I can consume content, make purchases and chat with friends using a mobile phone, but I could not have written this post or done the research that went into it without a laptop and a fast Internet connection. High-speed connectivity to homes, schools, libraries, clinics, etc. will be in demand and narrow the qualitative digital divide.

Today's satellite company executives have different backgrounds than Teledesic's. Bill Gates had built a software company, Craig McCaw a cellular phone company and Boeing had experience with the technology of the time. I know nothing about Boeing's people today, but they have the benefit of the lessons the company has learned since Teledesic. OneWeb CEO Greg Wyler created a successful satellite-based company, O3b, which provides Internet service using a constellation of 12 mid-Earth orbit satellites. (O3b is now owned by SES).

Given his track record, Elon Musk deserves special mention. For a start, he is CEO and Lead Designer at SpaceX -- actively involved in engineering decisions -- Jobs and Wozniak in one package. He is also unfettered by outside investors, giving him more latitude and the ability to reassign personnel -- relatively free of shareholder's pressure to pursue their applications or maximize short-run profit. (Of course, he needs to make money in order to continue his business and invest in journeys to the Moon and Mars). While I hope the market is large and varied enough to support others, SpaceX will succeed if anyone does.

Unknowns and possible glitches

The above changes improve the outlook for today's satellite projects, but there is still uncertainty. For example, it is hard to know what the true capacity of these systems will be -- the number of users and the user mix they will support. They speak of serving individual users, local area networks like schools, libraries or public-access hotspots, Internet gateways, cell phone backhaul, ships, airplanes, automobiles, Internet of things (IoT) devices, etc. One can even imagine a space-based content-delivery network using solid-state drives.

An Internet gateway requires more bandwidth than an IoT sensor that is read every hour and, during the day, a school needs more bandwidth than a home. Will the companies end up with different customer mixes? For example, SpaceX has singled out long-haul Internet exchanges and home connectivity as target applications and, given Elon Musk's other interests, we can imagine him targeting autonomous vehicles and power distribution. OneWeb is also going after homes and buildings, but a key investor, Softbank, is focused on the Internet of things. OneWeb CEO Doug Wyler's first satellite service, O3b, serves mobile network operators, ship lines, governments, and enterprises. Perhaps OneWeb will focus on large customers.

A related question -- how will these companies handle sales and service? Will they have online sales, global offices, dealers, partner with ISPs and mobile companies, etc.? (OneWeb backer Softbank tried to merge with Intelsat, but the merger, which would have provided OneWeb with global offices, failed).

There are regulatory as well as technical challenges. The FCC has delayed acting on SpaceX's launch application pending negotiations on spectrum-sharing techniques between them and other satellite companies using the same range of radio frequencies. These will be global networks and they will have to satisfy the spectrum regulators of all nations they wish to serve as well as the International Telecommunication Union. At the very least, that means dealing with a lot of bureaucracy.

Standards also come to mind. SpaceX, OneWeb, Boeing, and others are working on methods of spectrum sharing. The Ethernet standard grew out of joint work by Intel, DEC and Xerox -- will SpaceX, OneWeb and Boeing create satellite spectrum standards one day?

Regulators also worry about falling debris. These satellites have a useful life of about five years after which they are de-orbited and burn up in the atmosphere. SpaceX hopes to launch over 4,000 satellites in their first constellation and they are in discussion with the FCC over the likelihood of human injury from falling debris. SpaceX eventually hopes to be able to economically recapture spent satellites using the BFR.

There are also political and cultural barriers. Will nations like China or North Korea allow citizens to install home ground stations and connect to the Internet? Will Saudi Arabia tolerate access to pornography? Cuba will not allow citizens to connect to expensive, slow geostationary satellites -- allowing connections to fast, cheap Internet satellites would require a political shift.

(That being said, Cuba would be a prime candidate for this service if the government would allow it. They have very little infrastructure, an educated population and since it is an island, the constellation "footprint" would not be densely populated).

There may also be a need for international anti-trust and consumer-protection regulation. What if one or all three of these companies succeed in establishing global networks that become monopolies or oligopolies in some nations or regions -- for example in central Africa. That could have significant consumer protection and political implications.

Teledesic planned to offer service through local service providers, but OneWeb and SpaceX plan to sell services and easily installed ground-stations directly to consumers. Will terrestrial Internet service providers fight them by lobbying governments and in court the way they have fought publicly-owned terrestrial Internet service providers?

Manufacturing, launching and operating these constellations seems like an immense engineering task. It is hard to imagine an organization capable of such a feat, but let's put these efforts in perspective. If humans can manufacture, launch, deploy and operate the James Webb Space Telescope or the flood-control and water-distribution system in ancient Petra, a constellation of "low-tech" satellites in low-Earth orbit seems like a piece of cake -- the question is whether the business model will work out.

Will these companies succeed?

Sorry to disappoint, but I don't know enough about their coverage and business models to say "yes" with certainty, but I am hopeful because smart, experienced business people, entrepreneurs, and investors are betting they will succeed.

More remarkable -- last year, Tom Wheeler, President Obama's FCC chairman, summed up a talk on the impact of improving technology, saying:
The pace of innovation is accelerating, and with new technological advances, satellites now have the opportunity to play a much more important role in bringing broadband to underserved and unserved areas around the world.
This year, Trump's FCC chairman Ajit Pai stated:
Today, the FCC updates the framework that will govern non-geostationary-satellite orbit satellite systems. And it’s high time: It’s been over a decade since we first adopted rules for these types of constellations. In the years since, innovation has brought exciting potential to connect consumers across the nation, especially in rural, remote, and tribal areas. The rules we adopt will promote the next generation of non-geostationary satellite systems, which could expand broadband access where it’s needed most.
Do Obama and Trump's appointees agree on anything other than this? Let's hope all the projects succeed and we have some decent competition.

Saturday, October 14, 2017

The BFR and its role in SpaceX's satellite Internet service

Elon Musk is CEO and Lead Designer at SpaceX.

SpaceX started with their Falcon 1 booster followed by several versions of the Falcon 9. The Falcon Heavy will fly later this year and the rocket that will take the first person to Mars is called, for now, the Big F***ing Rocket or BFR.

The Falcon1, Falcon 9, Falcon heavy and the BFR (source)

The 150-ton BFR payload will be 10 times that of the Falcon 9. It will have an have an extra landing-guidance engine for reliable reusability and SpaceX also expects to be able to soft-land and reuse the second-stage payload rocket as well as its protective nose cone, substantially reducing cost per launch. (Note that Boeing is also planning a Mars mission so they may be planning their own BFR).

The following is speculation, but I think the BFR will play a significant role the SpaceX satellite Internet service.

SpaceX applied to launch their 4,425 satellites in two phases -- an initial deployment of 1,600 satellites and a final deployment of 2,825. That is a lot of satellites and the FCC has required licensees to deploy their full constellations within six years of their grant, but last month they relaxed that constraint, establishing milestones of launching 50% of a constellation within six years and allowing another three years to complete the constellation. The FCC has delayed licensing SpaceX's plan until spectrum sharing agreements are reached by satellite operators, so the clock has not yet started running on their six and nine-year milestones.

SpaceX plans to send a BFR to Mars in December 2022, and they won't give me any details, but they will surely be used "locally" before that. They plan to begin launching operating Internet satellites in 2019 and those will be launched by Falcon 9 or Falcon Heavy rockets, but the BFR should be available to launch many of the planned 4,425 satellites before the FCC deadline and it will be used for replacement satellites when they are eventually required.

SpaceX estimates the satellite mass as 386 kg and the BFR can carry a 150-ton payload so, if they fit perfectly, a BFR could launch about 350 satellites at a time, but they won't fit perfectly, so let's say 300 per launch. SpaceX Senior Director Tom Ochinero says they will be capable of up to six launches per month. Using the BFR, 4,425 satellites in nine years sounds feasible and relatively cheap. (Elon Musk has estimated that future versions of the BFR may carry up to 1,000 tons).

The BFR may also play a role in debris mitigation. When they are taken out of operation, satellites are de-orbited and they burn up in the atmosphere, but there is some risk of debris hitting the Earth. Bloomberg reported that the FCC had challenged SpaceX's assessment of risk of human casualty from falling debris and SpaceX responded the following month. Recently two Senators have also asked the FCC to investigate the risk of collisions and debris.

The BFR may render the debate moot. In a recent presentation, Elon Musk speculated that the BFR might be used to capture orbiting satellites and return them to Earth, as illustrated here:

SpaceX hopes to recapture satellites in the future (source)

I will conclude with the following image that illustrates how the BFR got its name -- it is a BFR. If you are interested in the BFR and its role in Elon Musk's plan to colonize Mars, you should definitely read the post this illustration is taken from.

Still not sure how big it is? Check out this view of a BFR in Boston:

Update 10/20/2017

In a talk at the 2017 International Astronautical Congress in Australia last month, Elon Musk summarized SpaceX's technology progress, including 18 successful booster landings, and described the BFR design and economics and its applications -- launching satellites, shuttling to the International Space Station and travel to the Moon, Mars and between distant cities on Earth. It is a terrific talk, well worth watching:

On October 15, Musk followed up his talk with an “Ask Me Anything" (AMA) discussion about the BFR on Reddit. You can read a good summary of the AMA discussion, which includes video excerpts from the IAC talk and a concept video on terrestrial travel here. if you have time on your hands to geek out, you can see the entire AMA session here.

Last, but not least, SpaceX has posted a terrific 39-slide presentation on the project. It's a "Steve Jobs" kind of presentation -- long on images illustrating a concept and short on words. (I'm a big fan of that presentation style and try to force it on my beleaguered students). The presentation also includes links to a couple of animations and a video illustrating a hypothetical terrestrial travel scenario. Here are four of the slides to whet your appetite:

Update 12/9/2017

The much-delayed test launch of the Falcon Heavy rocket is scheduled for January 2018. The payload capability of the Falcon Heavy is about 2.5 times that of the Falcon 9 and around one-third of that of the BFR -- "well over 100,000 pounds to low-Earth orbit" according to Elon Musk.

(The Falcon Heavy test payload will be a Tesla Roadster destined for solar orbit -- a good publicity stunt and a near-permanent symbol of our transition to alternative energy sources).

The Falcon 9 will be used to launch SpaceX's first two prototype satellites next year, but in 2019, when SpaceX will begin launching operational Internet-service satellites, the Falcon Heavy will be available. The BFR will be available before the first Internet constellation is complete in 2024.

Update 12/20/2017

SpaceX has released photos of the first Falcon Heavy rocket. It is expected to launch next month, putting a Tesla Roadster in solar orbit. When asked why he wanted to put the car in orbit, Musk said he loves "the thought of a car drifting apparently endlessly through space and perhaps being discovered by an alien race millions of years in the future," and so do I. That reply is even cooler than Mallory saying he climbed Mount Everest "because it's there."

They hope to retrieve and reuse the three booster rockets.