Thursday, December 26, 2013

A Cuban approach to achieving Internet connectivity

(Spanish version)

Cuba's domestic Internet infrastructure is one of the worst in the world and the prospects for improvement are dim because of the U. S. embargo, a Cuban government policy of access control and scarcity, ETECSA's power, the lack of a trained, demanding technician and user base and a lack of capital. Can these obstacles be eliminated?

The embargo will eventually be dropped, and there are signs that that may be relatively soon. In the interim, China and others are willing to sell to and trade with Cuba.

Governmental control policy can change. When Cuba first joined the Interent, there was high level debate over the dictator's dilemma -- the perceived political and cultural threat of the Internet versus its value in improving people's lives and the economy. The decision was made to control the Internet and access to it, but that is not set in stone -- it can be reversed.

How about ETECSA? Is there any nation in which the incumbent telecommunication provider -- whether government owned or privately held -- has not acted in its self interest to the detriment of the people and economy? I suspect the answer to that question is "no." I have no knowledge of the current management of ETECSA, but I would be surprised if they were different than others.

ETECSA is jointly owned by the Ministry of Information and Communication and RAFIN, SA. The Ministry is of course part of the government and subject to political will -- policies and leaders can change.

RAFIN is a different matter. I don't know what their role is in the management of ETECSA. I don't even understand what the role of an "SA" is in a socialist nation. Where did they get the capital to purchase Telecom Italia's share of ETECSA? Who are the shareholders and investors? Do they share ETECSA profits and losses? Do they have a "seat on the board" -- a voice in picking executives and making policy decisions? I need the help of an economist here.

A trained, demanding technician and user base will come after connectivity becomes useful, widely available and affordable -- it will follow, not lead the path to a modern Internet.

That leaves the lack of capital. The Chinese took an active role in the financing and installation of the ALBA-1 undersea cable, and we speculated that they might also invest in complementary domestic infrastructure, but that has not happened.

The conventional wisdom from the World Bank or International Telecommunication Union is that the way to raise capital for connectivity is to privatize the telecommunication/ISP industry, and invite foreign investors to build infrastructure and compete on a level playing field watched over by a regulating agency -- privatization, regulation and competition (PCR).

Raúl Castro announced that they are working on a new foreign investment policy, which is of "singular importance to stimulate economic and social development of the country." The law is expected to be approved next March. It remains to be seen whether the new law and perceived demand would attract major investors, but even if they would, there is a problem with the PCR strategy -- it does not work well.

Many developing nations opted for PCR between 1991 and 2008:


In 2009 I looked at the data and concluded that "PCR has had little impact on the Internet during the last ten years in developed or developing nations." I have not updated that paper with subsequent data, but our experience in the US shows that private ownership of telecommunication service providers does not guarantee competition, efficiency and good service in spite of the good intentions of the regulators and congress.

We need a Cuban solution.

It would be great if Cuba could afford modern telecommunication infrastructure, with fiber to premises and backhaul for LTE mobile communications, but it cannot, so we need to think about cheaper interim approaches. The remainder of this post will speculate on one possibility -- a decentralized, multi-satellite policy.

Several years ago, I wrote a couple of articles (here and here) surveying wireless technologies for connectivity in developing nations -- tethered and untethered high-altitude platforms (HAPs), terrestrial wireless (WiMAX was a hope at the time), low-earth orbit (LEO) satellite constellations and very small aperture satellite terminals (VSATs).

Google is experimenting with HAPs, but there are no meaningful deployments. As far as I know, no one is studying LEO satellites and WiMAX has not developed as envisioned. At the time of those earlier articles, VSAT was the only option for connecting rural areas in nations like India, but VSAT ground stations were large, expensive and slow.


Since that time, technology has progressed and the consumer market for satellite connectivity has grown. U. S. providers HughesNet and Viasat have 1,398,000 subscribers between them. In spite of long latency times, I have had smooth video conversations with friends using home satellite dishes in rural Brazil and Chile. The antennae are small, costs are down and speeds are up.

What if Cuba were to encourage the use of these dishes rather than ban them?

Cuba has said they will authorize agents to sell telephone and Internet time. What if they were to expand the program to allow those agents to own and sell time and services using satellite Internet links, in the same way Grameen Phone ladies in Bangladesh bought mobile phones to resell call time.


Today, there are a few illegal satellite installations in Cuba. Imagine 1,000 legal satellite dishes dispersed throughout the island providing Internet access and VOIP calls (which are illegal today).

If that were to be considered, I imagine ETECSA would want to own the ground stations and set prices. That would insure profits and government control over Internet access, but it would be short sighted. Allowing the satellite operators to own their own equipment, would create a decentralized, self-organizing group of entrepreneurs who would bring effort and innovation to the project.

The situation in Cuba today is reminiscent of the Internet in the late 1980s in the U. S. TCP/IP had been invented and shown to be effective in the APRANet and CSNET. The potential of the network was obvious to those who had used it, but access was restricted to a few organizations and people.

In order to bring others online, The National Science Foundation established NSFNet. They contracted for a national backbone network and offered all U. S. colleges and universities grants to cover the cost of a router and a connection to the backbone. They also offered connectivity to education and research networks in developing nations. When it was decommissioned in April 1995, NSFNet was the global backbone, linking 28,470 domestic and 22,296 foreign networks. (Note that Sprint, the developing nations connectivity provider, also provided connectivity to Cuba in spite of the embargo).

The entire NSFNET project cost the U. S. taxpayer $94.5 million -- a small investment with an inestimable return. Blanketing Cuba with small satellite dishes would have similar results.

The NSFnet investment was highly leveraged. While universities got free connections to the backbone, they were expected to provide access for faculty and students. Collectively, universities invested much more in their campus local area networks, training and staff than NSF did in NSFnet. The decentralized approach and the end-end network architecture pushed both capital formation and innovation to the edge of the network where there were eager investors and entrepreneurs.

What would be the role of the Cuban government in a decentralized satellite access world? Their most important task would be capacity planning and negotiating with satellite communication companies for bandwidth. They would also specify, evaluate and purchase ground station equipment (some of which could be manufactured on the island).

They should also take the lead in developing software for efficient offline operation with automatic compression and data transfer when the user goes online. That software would be useful in any limited bandwidth nation, not only Cuba. Necessity being the mother of invention, we might even see some novel solutions for busy executives travelling in "airplane mode."

The government should also support the satellite operators by offering loans to help with initial equipment costs and by facilitating training and the sharing of experience and best practices. One can imagine a government run micro-finance bank offering loans and the government paying the overhead costs for a satellite operator's association. As was the case with NSFNet, the government could phase out of some of this activity once the network was stable and self-sustaining.

Of course the satellite system is an interim step -- in the long run, it will be phased out in favor of modern fiber infrastructure. The satellite system would pave the way to that goal by building user skill and demand. The satellite links would also guide the government in allocating scarce fiber resources -- high demand areas would be connected before others. (Google followed a similar strategy in prioritizing neighborhoods when rolling out their gigabit network in Kansas City -- areas with many committed subscribers were the first to be connected).

Note that I have suggested the government be responsible for a fiber backbone, but not for providing Internet service. They should view the backbone as they view highways -- providing infrastructure for use by independently owned trucks, buses and cars. China followed a similar Internet rollout strategy, with government organizations building backbone networks that, by the end 1999, were being used by over 500 Internet service providers.

Recall that the NSFNet universities provided their own local area networks. One can also imagine pueblo or ciudad-area networks linking the ground stations in a town. As with NSF, the design and investment in any such networks should be local. In this case, I am reminded of the home-made TV distribution networks, in which people would install their own coaxial cable to connect homes and other locations to a central ground station.


At the start of this post, I listed hurdles along the road to Cuban connectivity. I have outlined a low-cost, bootstrap proposal for connectivity that does not require foreign investment.

That leaves the political hurdles. Maybe there is hope. As noted above, the U. S. has signaled a desire for political change and Raúl Castro has admonished Cubans to embrace economic reforms "without haste, but without pause."

More specific encouragement comes from First Vice-President Miguel Díaz-Canel, who has stated that "Today, with the development of information technologies, of social networks, of computing and the Internet, prohibiting something is almost a chimera, impossible ... makes no sense ... We must constantly be in dialogue."

I recognize the irony in proposing that the government embrace a technology that led to the imprisonment of Alan Gross and others. Reversing the law on satellite communication would require political courage, but it would also provide the government a powerful argument against the charges leveled against them and they would be pursuing a Cuban solution -- one in which the Internet is operated as a service to the people and society, not the government or telecommunication companies.
-----
Update 1/5/2014

A person commenting on this post argued that U. S. policy would not have to be changed for this proposal to succeed -- he suggested that the policy changes announced by the administration in April 2009 cleared the way for sales of satellite Internet service.

The administration fact sheet on Reaching out to the Cuban People authorizes fiber-optic cable and satellite telecommunications facilities linking the United States and Cuba. It goes on to explicitly allow satellite radio and television service, but does not mention Internet service.

I sent an inquiry to the Treasury Department asking if a satellite Internet provider would be able to get a license to serve a Cuban account. A spokesman replied that he would find out and let me know.
-----

Update 1/17/2014

Satellite ISP IPSTAR says they have connected over 26,000 schools in Thailand, allowing more than 2,000,000 students access to online learning materials and IP-based applications. They downlink to LANs in schools and learning "cafes" and focus on delivering teaching matrial. This program appears to be relatively centralized and narrowly focused, but it is an example of a government sponsored satellite connectivity project.

Here is a short IPSTAR video on education and other applications:


----

Update 2/10/2014

As noted above, there are political obstacles to this proposal in both Cuba and the U. S. I asked the Office of Foreign Assets Control (OFAC) of the U. S. Treasury Department, which oversees our Cuba trade policy, about this proposal.

OFAC's Cuban Assets Control Regulations policy regarding the Internet is as follows:

§515.578 Exportation of certain services incident to Internet-based communications.

(a) Except as provided in paragraph (b) of this section, the exportation from the United States or by persons subject to U.S. jurisdiction to persons in Cuba of services incident to the exchange of personal communications over the Internet, such as instant messaging, chat and email, social networking, sharing of photos and movies, web browsing, and blogging, is authorized, provided that such services are publicly available at no cost to the user.

(b) This section does not authorize:

(1) The direct or indirect exportation of services with knowledge or reason to know that such services are intended for a prohibited official of the Government of Cuba, as defined in §515.337 of this part, or a prohibited member of the Cuban Communist Party, as defined in §515.338 of this part.

(2) The direct or indirect exportation of Internet connectivity services or telecommunications transmission facilities (such as satellite links or dedicated lines).

Note to §515.578(b)(2): For general licenses related to the provision of telecommunications services between the United States and Cuba and contracts for telecommunications services provided to particular individuals in Cuba, see §515.542(b) and §515.542(c), respectively, of this part. For a general license and a statement of specific licensing policy related to the establishment of telecommunications facilities linking the United States or third countries and Cuba, see §515.542(d) of this part.

(3) The direct or indirect exportation of web-hosting services that are for purposes other than personal communications (e.g., web-hosting services for commercial endeavors) or of domain name registration services.

(4) The direct or indirect exportation of any items to Cuba.

Note to §515.578(b)(4): For the rules related to transactions ordinarily incident to the exportation or reexportation of items, including software, to Cuba, see §§515.533 and 515.559 of this part.

(c) Specific licenses may be issued on a case-by-case basis for the exportation of other services incident to the sharing of information over the Internet.

The policy disallows satellite Internet connectivity services, which I have proposed here, but it does allow for specific licences on a case by case basis. When I asked about that, a spokesperson stated "Off the record, I just don't think our licensing policy has extended that far."

I checked with a second expert, anonymous source who disagreed, stating that a license probably would be granted and that an explicit change in policy was in fact under consideration.

He also pointed out that the biggest sticking point might be the issue of garnishment/attachment -- U. S. companies are afraid that by entering into business with Cuba/ETECSA, they would open themselves to lawsuits in the U. S. by Cuban-Americans trying to recuperate damages for expropriations by the Cuban government. This roadblock might require Congressional legislation to “protect” U. S. companies from such suits.

The bottom line from the U. S. side seems to be that there are obstacles, but there seems to be a desire to overcome them -- that would leave the ball in the Cuban government's court.

-----
Update 8/22/2014

MIT Media Lab founder Nichlas Negropont gave a TED talk summrizing his work over the last 30 years. He concludes with his plan for the future -- using stationary satellites to connect the "last billion" -- the poorest, rural people -- to the Internet. (That part of his talk begins at 17:05). He mentions that he has a partner in this project -- might it be Google?


Here is what he had to say:
And so my plan, and unfortunately I haven't been able to get my partners at this point to let me announce them, but is to do this with a stationary satellite. There are many reasons that stationary satellites aren't the best things, but there are a lot of reasons why they are, and for two billion dollars, you can connect a lot more than 100 million people, but the reason I picked two, and I will leave this as my last slide, is two billion dollars is what we were spending in Afghanistan every week. So surely if we can connect Africa and the last billion people for numbers like that, we should be doing it.