Since ignorance is bliss, there are even
larger threats that go unchecked. Embedded systems are a sure candidate of the threat of
the millennium challenge. In fact, the threat is said to be at least four times larger
than that of Y2K problem in mainframe applications. But sadly, the embedded system Y2K
awareness is where mainframe Y2K awareness was two years ago. And one doesn't have those
two years to go before the embedded systems crash.
Embedded systems sport electronic chips
which have been pre-programmed at the factory and are used in all devices that offer
electronic control. From personal use products to business equipment, from infrastructure
set-ups to automobiles...all have embedded electronic chips. While many of the systems are
either not date-sensitive or may not create significant problems leading to breakdown,
many others will present significant problems or at least inconvenience. Credit card
scanners, electronic locks, HVAC systems, elevators, telephone switches, and other major
systems, as well as smaller systems face the risk of failure.
Nature Of The Problem
The embedded systems issue is especially untrackable because of their peculiar nature.
First, there are so many chip manufacturers and so many embedded systems that use them.
One can safely say that it equals the million uses of electronics. Second, every design
engineer has worked hard to make sure that his design is unique and hence there is no
commonalty in design. Third, there is no standard and documented manufacturing process.
And lastly, one cannot test one typical system and then extrapolate the results to all
others from that manufacturer or installer. This is what differentiates the embedded
systems problem from the mainframe one.
color="#000000" size="2" face="Arial">There are three components of the PC that may get
affected with Y2K :
The software which is 'burned-in' to a ROM
chip, called BIOS.
The format of the information stored by the
BIOS in the battery backed-up RAM, called the CMOS.
The output of the Real-Time Clock (RTC)
chip which gives the contents of the calendar.
These apart, there is the microprocessor.
Intel's definition of 'Year 2000 Capable'
says:
An Intel product, when used in accordance
with its associated documentation, is 'Year 2000 Capable' when, upon installation, it
accurately stores, displays, processes, provides, and/or receives date data from, into,
and between the twentieth and twenty first centuries, including leap year calculations,
provided that all other technology used in combination with the said product properly
exchanges date data with it. The only proper method of accessing the date in systems with
Intel motherboards is indirectly from the RTC via the BIOS. The BIOS in Intel motherboard
and baseboard contain a century checking and maintenance feature that checks the last two
significant digits of the year stored in the RTC during each BIOS request.
These systems often require someone with
detailed product knowledge to identify the component and test it. In many instances, it is
only the vendor or the supplier who is in a position to know how to fix the embedded
system. Older equipment may no longer be supported by the OEM, and an upgrade or
replacement is required. Conducting an assessment or inventory of everything an
organization owns is not possible. So one has to perform triage.
There exists no standard methodologies to
handle the embedded systems problem since the awareness itself is quite low. Says Peter de
Jager, a Y2K consultant, "There is a methodology of sorts. One can list all
electronically controlled equipment. Then ask, if that fails, do I care? If yes, then what
is the best way to restore it. This way one can minimize the damage and have some form of
control over the embedded systems problem. But this is pure logic and hardly a
methodology."
Interestingly, in a US organization that
started testing its embedded systems for Y2K compliance, the systems that got the highest
ranking were those that dealt with communications-e.g. phone, LANs, digital phones,
pagers, and security systems. If these systems fail, then it surely would spell havoc
across the world.
In an area where awareness in the US is
quite low, an Indian organization has started tackling the embedded systems problem.
Hindustan Lever Ltd (HLL) has undertaken a major Y2K effort that spans the entire
operations including the process control mechanisms. According to Lalit Sawhney, Head
(IT), HLL, "The impact of embedded systems problem is unascertained and therefore
neglected. But could one imagine the petrochemical equipment, the distributed control
systems in the chemical industry, the telecom switches, the maintenance systems of
aircraft and the like, all malfunctioning from a particular date? Even if it is not a
question of competitive advantage, it is a matter of unhindered business operation."
There are some pertinent questions that an
organization has to ask itself. What is the probability that there is an embedded system
problem? What is the level of business risk from embedded system failures? What level of
business risk is there in not knowing about potential failures of suppliers, customers,
and service providers? Some dismiss the potential of these problems with offhand comments
about shavers, VCRs, and TVs. There are problems in these systems, but such problems are
symptomatic of much more serious challanges in the fabric of our lives. We are surrounded
and maintained by embedded systems and even if 1 percent of those billions fail, the
impact will be significant.