PRASANTO K ROY, a FORMER
ASSISTANT EDITOR of DATAQUEST,
is EDITOR, COMPUTERs@HOME and PC QUEST
A planet with ten billion
embedded micros, five times more warfare infotech than weapons, a million terabits of
bandwidth. How did we get here?
hspace="2" width="180" height="239">I
face="Arial">n the beginning was the game.
In the last year of the 20th century, when production chips first crossed the GHz barrier,
it had seemed almost bizarre that it happened not in a massively-parallel supercomputer,
but in a $ 299 TV game console.
Last year, 2012-13, Indian users spent $ 6
billion on game software. But even in the nineties, the real high-end stuff was driven by
gaming, and it was among the most deadly-serious of things that happened to commercial
infotech.
It wasn't productivity apps and business
software that triggered India's PC explosion at the turn of the century, it was kiddie
games and adult games, and what were then ultra-powerful home theaters built around cheap
GHz RISC chips, megadisplays, and Bose audio systems. The PC-entertainment convergence had
begun then, somewhat later than predicted. Gaming was 40 percent of India's software sales
then, and about a quarter of its software exports-India finally broke into shrink-wrapped
stuff with games. Bandwidth was still scarce, and access to the early Internet was ISDN
for most people, 56 K for the low-end, and cable for the serious gamers.
The gamers always drove the specs. As we
crossed into this century, desktop and NCs were inadequate for gaming. They struggled even
with bloated office software.
Interesting how the performance and
reliability curve changed over time. In the last 10 years of the 20th century, it had
actually declined, as conventional chips kept pushing clock speeds but without keeping
pace with huge OSes and apps. Then, in the year 2004, came what is now accepted as a
landmark in the history of computing. That was when the IEEE Real-Time Performance
Initiative (RTPI) committee defined the acceptable lower limits of user interface and
system response, driven largely by gaming industry initiatives.
RTPI 12.1 said that a system should respond
completely to any user input or interrupt within 10 ms of the event, as long as less than
5 million cumulative integer and floating-point operations were involved, and as long as
the output didn't need to be time-spread (such as a movie). But Revision 12.6 added the
requirement that the application or OS could not terminate without completely committing
all data and instructions to disk. Post such termination, the system must recover
completely within 100 ms with no loss of data, instructions, or buffered user input.
Effectively, they were saying: your system should be usable, fast, and crash-proof.
Suddenly, military-class specs were brought into office and home PCs. Or, more precisely,
gaming specs.
This was challenged five times by
Microsoft, three times by Intel, and four times by Compaq, but the committee stuck to its
guns: you could do what you liked, but if you wanted a RTPI 12.1 compliance stamp, you
conformed. It took two years for the major software and system vendors to respond with
compliant systems.
A parallel IEEE committee spent two years
defining a standard interface specification, covering conventional, speech and video input
and output. Finally, this was included as a usability extension to the same RTPI 12.1.
Speech and video recognition because viable, appeared on public interface terminals
everywhere. Each year experts predicted the death of the keyboard and mouse, and each year
the world made more of them. Till today, there's no sign of the century-old QWERTY
keyboard being threatened by speech and vision interfaces, especially in offices, where
the cubicles have gotten smaller and closer together, and there's nothing as valuable as
privacy.
But finally, computers and their software
and interfaces begun to get really usable by public utility users, housewives, fathers,
hourly-wages labor-and not just by kids and a few million Internet-privileged gamers. By
the year 2008, there were 3 billion computers, one for every three people on the planet.
In addition, there were at least 10 billion embedded computers in every electric and
electronic device ever made.
How that really happened is another story.
I
color="#000000" size="2" face="Arial">n the year 2002, the international federation of
central reserve banks got together and asked the heavily funded three-year-old Cybercrime
division of Interpol for help in killing salami-slicing.
Salami-slicing was an old cracker crime.
Once confined to disgruntled and tech-savvy bank employees, it became popular among remote
crackers. It was simple: take 1 million savings accounts, and round down each individual
balance to the nearest rupee. Transfer those paise or cents in one sweep into a specific
bank account. It was rare for account holders to notice the difference, and almost no one
would complain about 63 paise becoming 00 paise.
In the early years of widespread cybercrime
at the turn of the century, the media continued to report it with a mix of amused
indulgence, hype, and sensationalism. It clearly considered online pornography the more
serious of the cyber issues. Then came the 2004 Cyberkiller, and a wave of original and
copycat crimes across the world that relegated salami-slicing to the inside pages of web
news services.
The Cyberkiller struck first in Germany.
Knoll Pharma's extranet on Knollstrasse in Ludwigshafen was penetrated on June 22, 2004.
The firewall retained some traces of the crack, and IS managers took the company's finance
subnet offline and took it apart. They found nothing. They didn't look elsewhere. That's
how they missed a soft agent left behind on an old NT6 server linked to the main CIM
workstation network. The worm stayed dormant for 10 days, then activated, making a few
small changes to the formulation for aspirin and a line of sleeping pills. The CIM system
obediently transmitted the edited formulation to the manufacturing control computers.
Another five days, and the altered drugs were out across the European Community (EC).
The first two deaths were reported from
Paris, the next one from Berlin. Six casualties and 24 hours later, the problem was linked
to the drugs. Another 12 hours passed before it was traced to the German source, during
which time there were two similar deaths in London-and the victims had taken aspirin
tablets not from Knoll but from Boots, UK. In another hour CNN Online kicked in with the
story, followed by Reuters and AP. There was panic, and all drugstores were closed around
the EC. The US FDA ordered all recent European drugs destroyed. The first recorded case of
cyberterrorism claimed 14 lives and cost Europe and others about $ 40 billion.
The Cyberkiller struck three more times in
four months before the Interpol traced her to a suburb in Sydney, Australia. The
23-year-old Stanford graduate, who turned out to have several petty phone-phreaking
convictions, was using a GPS2 pocket-comm device. She didn't know that it had an invisible
tracer ID pattern put in on a US Defense Department request to the manufacturer. The
tracer ID simply embedded the device code alongwith the GPS-derived location of the
device, accurate to within 1 meter on the planet surface and within 10 cm altitude error,
and attached it to any voice or data transmission's header.
The events had several effects. The
Cyberkiller got a conviction and death sentence in three months in an international court.
Interpol's CyberCrime division got a $ 100 billion annual funding for counter-terrorism
surveillance, equipment, and research. Even the NCRB in India beefed up its cyber crimes
department drastically, as did central law-enforcement agencies across the world. The
media turned away from its indulgent reporting of delinquent whizkids and reported on the
events with restrained revulsion and a very uncharacteristic maturity.
The CyberKiller left a lasting impression
on the infotech industry long after she was killed in an electric chair in the year 2005.
Her most significant 'contribution' to it was the year 2006 global agreement on equipment
tracing. A committee involving all members of the UN Security Council, including India,
and several standard bodies including the IEEE, were formed. In record time it formulated,
ratified, and mandated a standard for embedding tracer codes in all equipment that could
communicate.
There had been four proposals for this
tracer signature system. The first was from Microsoft, which had in place just such a
64-bit code for tracing pirated software, and which offered it for 2 cents per license.
The code was embedded onto streaming video and Microsoft offered $ 10 million to anyone
who could crack, block, or edit it. A Finnish group of four teenagers calling themselves
the GSM Phreakers Club won that money in 48 hours. Finally, an alternative proposal from
the Linux User Group, which involved a long-range non-repetitive data pattern modification
rather than a header code, won. The tracer signature format included a device ID, GMT
real-time stamp, and a configurable location stamp. The location was to be picked up from
either the on-board GPS information, if the device had a GPS receiver built in, or from
the electricity supply. By the year 2007, all electricity supply around the world was
encoded with local-time and location information added on at the local distribution
points. The embedded tracer hardware, if it did not have an on-board GPS, would simply
pick up the rough location information. Battery-run equipment picked up the information
easily out of stray RF from the mains; a year later, all portable devices mandatorily
included a GPS device, adding $ 20 to all their pricetags.
During the year 2008, there was the first
serious cyberwar attack, as the media delighted in calling it. Over six days, crackers
attacked US networks and services. The first was on two Internet-phone switching network
hubs in New York and New Jersey, taking down for 10 minutes all communications on the US
East coast. The second was on the Pentagon. The intruder came in through an official's
secretary's computer, penetrated a missile control network, and played with it for 24
hours before he was traced to a building in downtown Bangkok and caught. The US military
announced the break-in much later, including the fact that the vulnerable missile control
network had been an elaborate decoy system, going all the way to controlling dud silos.
This was part of a cyber-defense project started in 1997 and turned into an almost
foolproof system over 11 years and many minor attacks.
The third attack was on a network belonging
to Johnson & Johnson, and the intruder made changes to a database which controlled the
manufacture of baby food formula. But this was an offline database, and when the system's
replication kicked in at midnight, the CIM computers registered the change, but reported
it to the officer on duty. The manufacture was interrupted, without any of the alterations
executed. The changes would have made the baby food toxic.
The fourth attack was almost simultaneous
on four airports: Gatwick in London; Frankfurt Main; Washington Dulles; and Schipol,
Amsterdam. It took out the air traffic control computers. However, the intruders did not
try subtle, sophisticated changes; they simply brought the computers down. Security
systems kicked in within seconds in all four places, isolating and bringing redundant
systems online in minutes. Emergency procedures put approaching aircraft in holding
patterns or deflected them away, and on-board collision avoidance systems overrode manual
controls and took over the avionics. There were no crashes or mid-air collisions.
Only a few of the intruders could actually
be traced. Most were using older battery-powered equipment without GPS2 tracers, and
short-burst sessions that interrupted the tracer patterns. A few were controlling
remotely-placed handheld PCs with other remote computers, so all that could be traced and
seized were the intermediate notebooks.
The media caught on quickly, despite a
military clampdown on information about the attacks. But it reacted with less panic than
it could have, and that contributed partly to the limited public panic.
What the media didn't catch onto then, but
the military did, was that each attack was far less severe than it could have been. The
attacks could have caused air crashes, severe power outages, or mass killing of infants.
It was as if the digital intruders were trying to prove a point, rather than actually
kill. It was only years later that media columnists speculated on whether the intruders
could have had business interests in the equipment manufacturing or software development
industries....
The US military recommended to the UN
Security Council and the Equipment Tracing committee that GPS3 units be implanted in all
electric devices that could communicate. This was quickly ratified and enforced: by then,
single-chip GPS3 receivers cost less than $ 1, and atomic clocks and transmissions had
progressed enough for 10 cm location accuracy for these tiny devices. The committee also
reviewed and modified the tracer pattern spec to a shorter burst train.
Today, five years later, every electronic
device made has tracer devices, whether or not it can talk. Most devices can:
refrigerators and washing machines routinely dial up for remote diagnostics, household
electricity meters transmit power-consumption data back to the company and billing data to
the household computers. Every automobile has a GPS device and interchanges position,
speed, and traffic conformance data to control units, apart from the onboard computers
that control most of the car's subsystems. The four major manufacturers of GPS equipment
total up $ 260 billion in revenues.
But the most dramatic impact of all this
has been on the military. From 8 percent of the world's military budgets allocated to
infotech 15 years ago, today over 75 percent goes to information systems, digital
counter-invasion hardware and software, network surveillance and monitoring, and redundant
systems.
This has been the most significant
influence on infotech development in the past 10 years. Sort of like back to its roots.
Quite a bit of infotech and Internet development had its origins in US military research,
35 years ago. Then the two moved apart, with the military largely a customer of infotech
for the next 20 years. Then, just into this century, it got apparent that wars were all
going to be digitally fought, and the military stepped up its involvement, research, and
use of infotech and networking in every country in the world. Investment in hardware and
ground troops declined. The infotech industry grew to seven times its size in these past
10 years, and India's software exports about 25 times.
There's an interesting footnote to all
this, and it goes back to gaming. Today, in the year 2013, the big suppliers of military
and commercial defense systems are also the world's biggest game software developers. In
the early nineties, computer games drew on warfare themes and simulations. Then they went
far ahead. Today, the military is largely dependant on technology developed for and
adapted from game machines. The beta testers of many of the simulation modules, and
strategy and counter-invasion software for military use, are teenagers paid on a part-time
basis at these companies. We've come around full circle.