Monday, March 31, 2008

BitTorrent and Comcast: Who Pays and How?

Back on Feb. 4 of this year, I noted how a group of Swedish software experts got in trouble for running a peer-to-peer system for distributing video content over the Internet. The claim made by the prosecutors was that most of the content was pirated. Well, that turned out to be a sign of things to come. For some months now, the major U. S. cable television and Internet network operator Comcast has been in a dispute with BitTorrent Inc., a firm that provides software allowing peer-to-peer sharing of video. And the outcome of the fight may affect how all of us pay for Internet services for years to come.

The first punch in the public fight came when BitTorrent accused Comcast of singling out users of BitTorrent's protocol for interference and interruptions when Comcast's network traffic got too heavy for comfort. At first Comcast denied any such discrimination, but later under pressure, spokesmen for the cable and network firm admitted that they were doing exactly that. Then the Federal Communications Commission got involved and has held public hearings about the matter. On Mar. 27 (last Thursday), Comcast announced that it was making a number of changes that will both eliminate the discriminatory network measures against BitTorrent users and should make improvements in everyone's service through increased software and hardware efficiency and investment. But that hasn't stopped the FCC from announcing another hearing set for Apr. 17 at Stanford University in the heart of Silicon Valley, where I'm sure they will find people with an abundance of opinions on both sides.

What is BitTorrent and how does it work? You may recall the flaps about peer-to-peer sharing of audio files over the Internet a few years ago. BitTorrent's protocol also uses the fact that a file that one person wants is usually stored on thousands of other computers on the network. But video files are thousands of times bigger than audio files, especially if we're talking about HD video, which is becoming increasingly popular. The process of getting only one source computer to send a gigabyte-size file (1,000,000,000 bytes) over the Internet to another computer is tedious, error-prone, and takes a long time. So BitTorrent draws upon many of the other computers that have the file in question and gets them to cooperate by sending different pieces of the file to the target computer. Somehow the software coordinates all this confusion of activity, and the end result to the user is that he or she gets the desired file a lot faster than if only two computers were involved.

But as with so many things, what's good for the individual may not bode well for the group. Comcast and other network service providers estimate that because of BitTorrent's popularity, as much as half of all Internet traffic at certain times consists of peer-to-peer file sharing of this type. Comcast has defended its actions against BitTorrent protocols simply as their attempt to manage their limited network capacity fairly so that other customers were not left out in the cold with impaired service.

The word "fairly" means ethics has come into the picture. This ethical question arises from a tension that was born with the Internet some two decades ago, a tension between two competing philosophies.

Call the first the egalitarian-vision philosophy: the idea that information should be free, all Internet users should have the same privileges and access, and that such ideas should be built into the technical machinery of the Internet. The founders and early users of the Internet were imbued with this philosophy, and its legacy lives on in the basic structure of Internet protocols.

The second philosophy is the commercial free-enterprise notion that the Internet is a means to make money, and you should charge whatever the traffic will bear. It was years before anyone figured out how to make money with the Internet, but with the coming of Google I think it is fair to say that some people, anyway, have managed to do that. This philosophy sees the market as the best arbiter of resource distribution and even matters of fairness. Although there are now a few coarse-grained ways of charging people who want faster Internet service more money, hardly anyone pays any surcharge that depends on how much you actually use the thing. That is, if you ask your service provider for high-speed Internet service, you get a monthly bill that's the same whether you never touched your computer that month, or whether you downloaded seventeen movies in ten days using BitTorrent.

The network operators argue, and with some merit, that if five percent of their customers tie up half the resources of the entire network, it is not fair to the other 95% who pay just as much but have their service degraded by the overcrowding due to BitTorrent traffic. One alternative that Time Warner Cable is reported to be trying out in Beaumont, Texas on a trial basis is "metered" Internet use. That is, if you use more than a certain bandwidth-time product, let's call it, then you pay an extra fee. Metered use flies in the face of decades of Internet tradition and egalitarian philosophy, but if such distortions of the market as those caused by BitTorrent users continue, something will have to change, and the network companies may resort to metering on a wider scale.

A curious analogy to what is happening now with BitTorrent and Comcast went on for over a century in New York City. Until the late 1980s, residential users of the Big Apple's water supply had no meters—they just paid a flat monthly fee. You can imagine how this affected the way people used water. Finally, meters were installed, and the city as a whole used 28% less water in 2006 than it did in 1979. The Internet isn't water, but like water, it is not an infinite resource, and we may have to start paying by the drink if we don't want the whole thing to break down.

Sources: Bob Fernandez of the Philadelphia Inquirer has reported extensively on the BitTorrent-Comcast dispute, and I used his articles published on Mar. 23 (http://www.philly.com/philly/business/20080323_Online_Video__Data_Tidal_Wave_.html) and Mar. 27 (http://www.philly.com/philly/business/20080327_Comcast_agreement_in_dispute_with_BitTorrent.html). The statistic about New York City water use came from the Wikipedia article "Environmental issues in New York City."

Monday, March 24, 2008

Sustainable—But At What Cost?

I read a lot of discussion these days about "sustainability," "sustainable engineering," "sustainable agriculture," and so on. Sustainability, we are told, is the key to solving everything from global warming to finding world peace. What exactly is sustainability, and what are its implications?

One of the most obvious features of today's technological economy is not sustainable: the use of fossil fuels, which means mainly oil, natural gas, and coal. However these resources were formed (and there is still a good bit of debate about that), everybody agrees it took millions of years, and we stand a fair chance of running through them in a good deal less than 0.1% of that time, say a few hundred years. So the use of fossil fuels for energy is not sustainable.

So what? If you look around for anything at all, technological or not, which has turned out to be truly sustainable over recorded history, the list is fairly short. Things like the practice of begetting and raising families, farming, the life of some cities (e. g. Damascus, which is one of the oldest cities on earth), and even a few (very few) business firms have gone on for hundreds of years or more, and show no sign of disappearing because of lack of resources. I could add the professions of doctors and lawyers, and let's not forget taxes, but not governments that levy taxes—the habit endures even though the taxing entities don't.

The proponents of sustainability want basically everything we do to be a part of that kind of list—a list of things which have long traditions going back over many cultures and governments into the past.

In an article in the current issue of The New Atlantis, Yuval Levin makes the point that certain ideas vigorously promoted by political liberals in the U. S. are actually quite conservative. Sustainability, if successfully implemented, fits right into this pattern. If all social activities, technological and otherwise, were sustainable in the sense that liberals usually mean, the activities would go on and on without having to end because of physical limitations. While certain features might change, the physical resources needed would be either renewable or permanent.

Now that is a very conservative picture, meaning that the physical essentials of technology would not change. If new materials were invented that required using something that couldn't be recycled and reused–then they wouldn't be sustainable, and you couldn't use them. Everything would be recycled, with energy coming only from the sun. (Strictly speaking, even the sun isn't sustainable, although we can count on it shining for a few more million years.)

What if we went to such a totally sustainable economy? Some things wouldn't change much at all. Most steel is now made from recycled scrap, for instance, so that wouldn't be much of a problem.

But what about concrete? I have toyed with the idea of recycling concrete, because as far as I know, you could apply enough heat to it, drive off the water, and get back the calcium silicate that was in the original Portland cement. The trouble with this is, it would be vastly more expensive (and energy-intensive) to make cement from recycled concrete (laboriously hauled back from wherever it was poured to the recycling plant where huge amounts of energy would be required), than it would be simply to dig up some more limestone and sand from the ground. Ah, but limestone and sand are not renewable resources. Yes, there is enough limestone and sand to last us a long time, but if you're going to be a sustainable absolutist, you can't use anything that isn't recycled or in principle, recyclable.

I'm pushing this idea to the limits to make a point, but the point is a valid one. Namely, some things are more easily sustainable than others, and it simply doesn't make sense to hold sustainability up as a practical goal for every technological field, unless we are willing to make some very weird and silly changes in the way we do things.

While I was on vacation last week, I toured Indian City U. S. A. outside Anadarko, Oklahoma. It's a sort of outdoor museum where seven different kinds of Native American dwellings have been constructed and preserved. It was pouring rain at the time, but that didn't stop our guide from pointing out the different features of the various structures which were, of course, made from all-natural materials: tree trunks, mud, grass, and so on. Native Americans were the first recyclers, he said, since when they were finished with a structure they just abandoned it and let it return to Nature.

Though I didn't say anything at the time, I had a big "Yes, but. . . " in mind. Although estimates of how many people lived in what is now called North and South America before 1492 vary from 8 million to over 100 million, the figure is certainly less than the approximately 900 million people that the New World harbors today. And the Americas are some of the least densely populated regions of the developed world. If we all went back to living the way the first Native Americans did, there is no way that we would all be able to survive, even if we all suddenly acquired the hunting, gathering, and rudimentary agricultural skills necessary for such a life. And if we managed somehow to eke out a living, few of us would enjoy rising at dawn, doing back-breaking manual labor all day, and retiring at dusk only to do it all over again the next morning.

The only time when something like this has been tried on a massive scale recently was the Great Cultural Revolution under Mao Tse-tung in the Peoples' Republic of China, from 1968 to 1976. Millions of intellectuals and other suspicious persons, including most of the faculty members at all Chinese universities, were summarily hauled off to the countryside for a little bucolic "re-education" that lasted seven or eight years. I have known citizens of that country who lived through that period, and they tell me that it set back their lives a decade or more, and the progress of the country by a generation. But it was certainly sustainable, in the sense that they were still living and probably consuming fewer resources than they would have in the cities.

Few if any of the proponents of sustainability have in mind a radical, total shift to something like that. Or if they do, they're not talking about it openly. I favor a reasoned, appropriate move toward more nearly sustainable technology when it makes economic sense, when its adoption won't cause undue suffering or disruption, and when it leads to more human thriving than formerly. But a draconian swift transition to a totally sustainable economy would be in most respects indistinguishable from a worldwide depression. And I hope we don't get to that point any time soon.

Sources: Yuval Levin's article "Science and the Left" appears in the Winter 2008 edition of The New Atlantis.

Saturday, March 15, 2008

Robot Rats and SARs for PEPs

Sometimes things happen fast in politics. On Sunday morning, March 9, Eliot Spitzer woke up to the beginning of his 63rd week in office as Governor of New York State, an office which served as a stepping stone to the White House for his predecessors Theodore and Franklin D. Roosevelt. He had an apparently unstained reputation for fighting corruption in high places, which he had earned during his seven years as New York State's Attorney General, going after everything from Enron-type financial scandals to prostitution rings.

Two days from now—on Monday, March 17—he will hand over the keys of office and become Private Citizen Spitzer. Earlier this week, the New York Times revealed that Spitzer had been a customer of a prostitution ring that was under federal investigation. This evidence was revealed by a computer scan of Spitzer's banking transactions—a robot rat, if you will. The political firestorm that the news report touched off must have convinced him that trying to stay in office was an exercise in futility. On March 12, he announced that he was resigning. Ironies abound in a situation like this, but an ironic twist of special interest to the technical community is that Spitzer was caught by software that he had himself encouraged banks to use during his years as Attorney General. How did it work?

Banks have ethical obligations both to their customers and to the governments in whose jurisdictions they operate. Customers expect banks to keep their collective mouths shut about private financial matters, and by and large, banks are pretty good at doing this. But law enforcement officials realized long ago that banks are where the money is, including illegally gotten gains from enterprises such as drug dealing and prostitution. That is why in 1970, Congress passed the Bank Secrecy Act. This act is why you have to fill out a form with some identifying information any time you engage your bank in a single cash transaction of more than $10,000.

Criminals are as adaptable as anybody, and soon they learned not to trip that $10,000 wire by breaking up transactions into smaller amounts. To plug this leak in the dike, Congress enacted the Money Laundering Control Act of 1986. Besides asking banks to report any transactions over $5,000 that looked like they were evasions of the $10,000 limit, it removed liability for over-reporting. This meant that if you got annoyed at being called by the FBI for a series of legitimate but large financial transactions, you could no longer sue your bank for falsely tattling on you.

As time went on, $5,000 became less and less money in real terms, meaning that without doing a thing, Congress gradually lowered the threshold on what banks had to report. After a few banks got in trouble for under-reporting and computerized banking became nearly universal, the banks had the bright idea of just reporting everything automatically that looked suspicious. But first they had to tell the computers what "looking suspicious" meant.

One factor they loaded into their software, believe it or not, was the degree to which their customers are "politically exposed persons" (PEPs for short). If you are a governor, senator, UN delegate, or other personage whose position makes you more likely either to be the victim of a corrupt action (e. g. blackmail) or perhaps the perpetrator, you get a high PEP rating, and the threshold for making the computer spit out summaries of fishy-looking activity is accordingly set very low. Spitzer, needless to say, was a PEP, and when several large transactions to one firm showed up on a report, the bank decided to file a Suspicious Activity Report (SAR, for short) with the IRS.

At this point, humans got involved, but they could not have done their jobs without the aid of large software programs that inspect millions if not billions of transactions every year. Initially the investigators thought the governor might be the victim of blackmail, but when they found out the firm was a front for a prostitution ring, things took a different turn altogether.

Computers don't join political parties, but the people who program and operate them do. This story shows how technology can help law enforcement with investigations that in times past would have been impossible because of the sheer volume of data to inspect. Back in the days when the most advanced technology in a bank was the Friden calculating machine sitting on the comptroller's desk, a person's eyes were the only way to inspect records. That limited the nature and scope of investigations, although it also probably made things easier to do informally that were strictly against the law, as favors both to criminals and to policemen and detectives. Today, the same criteria can be applied impartially and exactly to millions of accounts, but at some point human judgment always comes into play. Once the computers provided the information to investigators, the investigators had to decide what to do with it.

And it was human judgment, however flawed, that made Governor Spitzer think that maybe he would escape detection of his expensive dalliances. Perhaps he was unconsciously hewing to an outmoded habit he developed before his own actions helped to tighten the screws on money launderers and others who do not care for banks to report their transactions to the government. Whatever the reason, this episode shows that the power to analyze large amounts of private computerized data can make or break very influential people. And without software engineers, no one would have that power.

Sources: A good summary of the laws and processes that led investigators to Spitzer's transactions is at http://firedoglake.com/2008/03/12/money-laundering-suspicious-activities-reports-and-structuring/. A Newsday account of how Spitzer's bank discovered the specific transactions is at http://www.newsday.com/news/local/state/ny-stspitzerbank0312,0,4637246.story.

Tuesday, March 11, 2008

Engineering the End of Malaria

In my Feb. 25 entry, I used the idea of wiping out malaria as an example of what might be done with "a few billion dollars" that would otherwise go toward dealing with global warming. I will admit that I simply pulled that number out of the air. Since then I have learned that while eliminating malaria is something that people as wealthy as Bill and Melinda Gates have tried to do, it is by no means a simple or straightforward task. But engineers may be able to help in some ways you wouldn't expect.

As you probably know, people contract malaria from the bite of a certain kind of mosquito that is infected with the protozoan parasite that causes the disease. The parasite hides inside liver cells or red blood cells in its human host, which is one reason that no one has devised an effective vaccine for the disease. Drugs are available to prevent it, but you have to take them all the time, sick or well, and such prophylactic treatment is too expensive for many residents of areas such as Africa where malaria is endemic. So many anti-malaria campaigns in the past have concentrated on eliminating the animal host: the anopheles mosquito that carries the malaria parasite.

The New York Times recently carried a report about whether malaria can be eliminated as smallpox has been. It seems that the consensus of public-health experts is that you can markedly reduce the incidence of malaria through spraying mosquito-infested areas with insecticide, but absolute elimination is an elusive goal at best. In Sri Lanka, for example, systematic spraying programs helped reduce the number of malaria cases from over a million in 1955 to only 18 in 1963. But the government cut back its programs, and malaria came back, reaching a level of over half a million cases in 1968. That lesson finally learned, Sri Lanka started spraying again and hasn't stopped, and the annual rate of malaria cases is now down to a few thousand.

At a 2007 malaria conference, Bill and Melinda Gates challenged public health leaders around the world to eradicate malaria altogether. Their foundation has already spent over a billion dollars fighting malaria, but clearly more than just money will be needed.

One commentator in Scientific American has pointed out that the free mosquito-net programs sponsored by many governments may not be as effective as they could be. Here is one area where engineers can get involved. The classic kind of mosquito net hangs from a string tied to the ceiling and drapes down to the edges of the mattress, protecting the sleeper from night-time mosquito bites, which is when the anopheles variety is active. This is fine as long as you have a mattress for the net to tuck under. But in thousands of villages where a mattress for every family member would be an unheard-of luxury, young children sleep on the ground. There are rectangular frame-type mosquito nets available that will work in this situation, but they aren't as convenient as the single-string type.

This little net problem is an example of how complex the malaria issue is. Even if engineers devised a new type of net that was ideal for the poorest residents, there are a lot of problems that remain. How do you get this net into the hands of those who can use it? How do you persuade them that using it will keep their children healthier? Who pays for all of this, especially if the new net costs more than the old ineffective ones?

In times past, some engineers would have said these issues were not engineering problems. But organizations like Engineers Without Borders (EWB) realize that the hardware or software part of a solution is only a part, and often not the most important part. An effective technical solution to any problem also has to factor in economics, motivation, distribution, education, and so on. EWB is an organization dedicated to providing engineering solutions for disadvantaged communities through sustainable engineering. Through many chapters at universities and colleges with engineering schools, they recruit volunteer students who get a holistic picture of not just a technical problem, but an entire culture and the cultural and social context of the problem as well. Though I never had such an experience in my student days, I think I might have been a very different kind of engineer if I had.

Only time will tell whether the wealth of the Gates Foundation, the ingenuity of engineers, medical researchers, public health officials, and the willingness of affected communities will converge to defeat the old tropical enemy, malaria. For the reasons I've discussed, it is a much harder task than the smallpox battle. But I wish the best for everyone involved.

Sources: The New York Times article on whether malaria can be defeated was carried in the Mar. 4, 2008 online edition at http://www.nytimes.com/2008/03/04/health/04mala.html. Scientific American's article on mosquito-net engineering appeared in the January 2008 issue, available online at http://www.sciam.com/article.cfm?id=a-better-mosquito-net. And Engineers Without Borders-International has a website at http://www.ewb-international.org.

Monday, March 03, 2008

Locked-In Profits or Service to the Downtrodden?

Suppose you're the wife of a man who got arrested in Oakland, California. You weren't with him at the time, and all you know is the bare fact that he was arrested. Until recently, your only alternative was to call the Alameda County public information number, work your way through a phone tree, and hope there would be a live person at the other end who could tell you something. Sometimes there was and sometimes there wasn't. But now, thanks to the initiative of some staff in the Alameda County Information Technology department, there is an Inmate Locator on the county's website. If you have the person's full name, or even if all you know is that they were booked in the last twenty-four hours, you can get online and see identifying information, the "custody status," and which jail they're in. Of course, you have to have a computer and a high-speed internet connection to do this efficiently, but doesn't everybody?

Despite the drawback of needing a computer to use it, this little advance in IT touches on a subject that I have seldom seen addressed in the engineering ethics literature. What special obligations or ethical issues are related to engineering as it applies to prisoners and jails? And in particular, what should we say about the recent trend toward privatization in U. S. prisons?

You may have read that the United States has the both the highest documented rate of incarceration in the world (over 700 per 100,000 population) and the largest absolute number of people behind bars (over 2 million, plus another 5 million or so on probation or parole). The reasons for this are worth going into, but for now let's just say they're a given. All these people have to be housed, fed, treated for medical conditions when necessary, shipped around, and maybe allowed some education and communication privileges. In addition, there are the families and friends of prisoners who have certain rights and privileges with regard to those behind bars. As the Alameda County IT folks have shown, engineering can benefit both the prisoners and their friends and relatives, in an entirely legal way (I'm not talking about high-tech jailbreaks here, which I suppose would be another way engineering could enter the picture).

I think it's significant that the people who came up with this idea were government employees (the article describing the system did not state otherwise). Along with the boom in prison populations has come a related boom in private prisons and companies that operate them. One of the largest, the Corrections Corporation of America, has gotten some coverage in this week's New Yorker magazine for its less-than-ideal operation of an illegal-immigrant holding facility outside Taylor, Texas, just up the road from my university here in San Marcos.

Privatization has been sold as a kind of universal solution to every government cost problem, but there are limits to what it can do. Somehow I suspect if Alameda County had outsourced its jail operations to a private firm, that firm would not have hired five web developers to come up with the Inmate Locator. Abuses can happen both in private and in public organizations, but the incentives are different.

As an employee of a state university, I view the advantages of well-run government-operated services as chiefly these: (1) Stability---the turnover in government employment is much lower than in comparable private operations; (2) Esprit de corps---in well-run government operations, a public-spiritedness can foster a selfless dedication to the needs of those serviced; (3) Relative lack of cost-squeezing pressures---assuming the management makes a good case to the appropriate legislature, expenditures can be planned and justified without concern that they will risk ending the whole enterprise if a lower bidder comes along.

I'm well aware that a critic could come along and turn each of those arguments on its head. Stability can mean that once a goof-off gets a government job, he's set for life. Private companies can develop esprit de corps too, and cost-squeezing pressures can happen in government as well as private industry.

But I would point out a philosophical difference between the two approaches. The bottom line of government service is just that: service. Ideally, the public servant is as dedicated to his or her clients as the nuns of centuries ago who founded and staffed the first hospitals. At least, there is no philosophical conflict between having a totally dedicated public servant and the overall goals of the organization.

With private companies, especially those which are joint-stock (publicly owned) firms, the fundamental philosophy is different. If a company doesn't make money for more than a certain length of time, it should disappear, and often does (despite evidence such as General Motors to the contrary). Companies can provide good services, but there is a built-in conflict between the ultimate raison d'etre of a company, which is making money for the owners, and service to its customers or clients, at least to the extent that improvements in the service or product make less profit available to the owners.

This is not to say that all corporate enterprise is morally suspect—absolutely not. But prisoners are a special kind of client, and are treated specially along with children, the elderly, and medical patients in a number of ethical contexts such as the rules for ethical conduct of research studies. Unlike a customer at a hardware store, if a prisoner doesn't like the service he's getting, he can't just walk away and go to another prison. I think that is the main reason why for nearly the entire history of prisons in the U. S., they have been exclusively a government-run operation. Maybe the government didn't do that good a job, but at least there was a way, in principle, for abuses in government-run prisons to be corrected through the democratic process. Private companies that run prisons can and do claim that vital information about their operations is a trade secret, and therefore not available for public access, at least not without a lengthy and often unsuccessful series of inquiries under the Freedom of Information Act. This kind of secrecy can hide abuses and wrongdoing that would be harder to hide in a public setting.

So what is the bottom line here? First, kudos to the IT folks in the Alameda County Sheriff's Office, who make it possible for the over 100 inmates booked each 24 hours to be found by their relatives or friends much more easily than before. Second, any time an engineer does something related to prisons or prisoners, he or she should remember that prisoners are not just any old client. They have special rights and privileges. Yes, many of them have done something wrong. But the fact that we are a country of laws means that we need to hold those laws in high regard, especially when we deal with people who may have broken them.

Sources: The article on Inmate Finder appeared in the online issue of the San Francisco Examiner for Mar. 3, 2008 at http://extra.examiner.com/linker/?url=http%3A%2F%2Fwww%2Einsidebayarea%2Ecom%2Fci%5F8435580%3Fsource%3Drss. The New Yorker article by Margaret Talbot on CCA's operation in Taylor is entitled "Lost Children," on p. 58 of the Mar. 3, 2008 edition. Statistics on U. S. prisons were found at the Wikipedia article "Prisons in the United States."