Transcript
Page 1: Windows-98, RIP and my Adrenaline Rush

Computers & Security, 17 (1998) 564-574

Security Views Dr. Bill Hancock, CISSP

Editor-in-Chief

Windows-98, RIP and My Adrenaline Rush

As some of you are aware, I have been an avid student

of the martial arts since age 4. I did rather well in

competition in the 70’s and 80’s at very large, interna-

tional tournaments. Its come in quite handy on a

couple of unfortunate times, but has always been part

of my life and has been there for me when I needed it

most.

What does this have to do with anything? Well, I still get an adrenaline rush when I step into a tournament

ring - even if I am just one of the other ‘old fossils’

who judge the efforts of the younger and more agile.

Of course, us ‘old fossils’ have a saying:

Youth md exuberance does riot stand a clzance against age arld deceit.. .

But, martial arts tournaments give me an adrenaline

rush. Probably always will.

Of course, other things do as well. One is when I come to a realization of how nasty a security problem

some ‘feature’ or another is in an operating system that I am fooling around with. With that, I recently had a small one that grew in scope when I put two-and-two

together and came up with a much larger number.

After recently purchasing a new Pentium II with all

the ‘bells and whistles’ on it, it came pre-installed with

the standard, buggy release of Windows-98. As the

proper geek that I am, once I was happy that every- thing was working, I parked myself on the Microsoft

Windows-98 CD-ROM that came with my system and started looking into items that are ‘optional’

installs such as theValuPack and other goodies that are

on the CD-ROM.

Whilst browsing the CD-ROM, I discovered that an

install existed for Routing Information Protocol (RIP).This was a bit of a surprise as making a desktop

system a router is not always a good idea for lots of

network control reasons and even more security

reasons. But, being curious as to whether it worked or not, I installed it and started playing with it.

I was quite successful in getting an ISP connection via

modem with PPP to route properly to an

Ethernet/802.3 connection on the system itself. This

meant that I could dial up an ISP and allow other nodes on my constructed testing LAN to communi-

cate with the ISP and, therefore, the Internet. A little

further extrapolation caused the revelation that any laptop running W98 with the optional RIP installed

would allow the laptop to ‘back door’ connect the laptop to the ISP on the modem side and the laptop to the corporate LAN on the Ethernet side of the connection. Hence an adrenaline rush - a secu- rity problem, and a pretty serious one, waiting to happen.

564 0167-4048/98$19.00 0 1998 Elsevier Science Ltd. All rights reserved

Page 2: Windows-98, RIP and my Adrenaline Rush

Computers & Security, Vol. 17, No. 7

For example, one of my customers has a site with 2000

telephone handsets on the property. They also have

over 950 analogue lines. Of those 950 analogue lines,

over 700 are modem connection lines for desktop

systems as the users got sick and tired of getting the IS

department to help them connect out to the world.

Using RIP with a W98 upgrade would now allow

these users to connect out to the world and also route

information from the local network to the Internet

and vice-versa.

Of course, using this type of network technique is not

new. Products such as WinGate have allowed this

capability on W95 for some time now. Other Network

Address Translation (NAT) products for W95 and MS-

DOS accomplish much the same thing. But, you must

purchase WinGate after the trial period for a small fee

and RIP comes included in the W98 distribution kit.

Further, you would also be required to do a little

snooping around the Net to figure out that there was

a WinGate in existence. Inclusion of RIP on the

distribution CD is much easier to discover and use.

I fully expect a lot of small businesses to implement

this optional component. 1 also fully expect large cor-

porate laptop users to do the same. In either situation,

compromise of the internal networks via unautho-

rized connectivity and access will become much more

commonplace. It also follows that if your company

does not have policies about what is allowed to be

connected to the corporate network and, importantly,

how, then be prepared to have your network compro-

mised by a desktop near you soon.

Stealth Probing of Internet-connected Sites

A recent series of long-term, low packet-count probing has been going on within various large

network sites belonging to the military and other large

sites. Low-bandwidth, or group, hacking involves

numerous hackers working together from different locations. Together, they intermittently send sets of IP packets against a network to test for vulnerabilities. Because the packets come from different hosts and at varying intervals, they come in, in effect, ‘under the

radar’ of most intrusion-detection applications currently on the market.

This type of attack has been rumored about for

several years, but it wasn’t until last month that it was

documented by the Shadow project of the US

Department of the Navy’s Surface Warfare Center.

With these new low-bandwidth attacks, hackers have

found a way to make the most obvious part of their

attacks - probing for vulnerabilities - virtually

undetectable. That frees them up to do the real

damage by racing through those holes to capture data

before they can be shut down. So far, there have been

three distinct patterns that have emerged:

Slow scans for machines and services: Attacker inter-

mittently checks for machines and services to develop

a picture of the target network. Once vulnerabilities

are mapped, attacker can go back through that hole.

Multisourced attack: Attacker tries to access or crash a

server, also known as denial of service, from multiple

points of origin.

Multisourced attacks to multiple targets: Attacker

dilutes the so-called attack density, making it look like

normal traffic that is converging on the same data.

So far there has been a lot of vendor posturing and

assurances of product upgrades coming out to fight

this new method of probing for weakness. We will see.

What is already known is that this type of probing

requires monitoring consisting of various database and

neural network methods as well as macho computing

hardware to handle the performance issues that are

sure to be an issue. This type of monitoring comes at

a price and it’s not going to be cheap to implement nor simple to solve.

Lotus Domino Security Flaws Redux Another security glitch on Lotus Development’s

Domino Web server may make it possible to view sensitive credit card, address, and phone data from the Web. The breach was reported last week by LOpht Heavy Industries, a group dedicated to security research.Their advisory is located on their web site at:

565

Page 3: Windows-98, RIP and my Adrenaline Rush

Security Vie ws/Dr. Bill Hancock

http://www.lOpht.com (that’s a zero, not an O).This

flaw could make Lotus Domino application-based

payment and client data available from a Web browser.

LOphtS website said it “received reports regarding a

vulnerability in some implementations of Domino-

based applications, which result in the Internet publi-

cation of sensitive information belonging to

customers of Lotus/IBM and their business partners.”

LOpht said Web browser users can access database

information simply by navigating to the payment entry part of a Domino site, then substituting ‘open’

after -.nsf database names in the URL. LOpht suggests

developers use reader and author names fields to

prevent unauthorized access to sensitive data. It also

suggests disallowing anonymous access to names.nsf,

catalog.nsf, log.nsf, domlog.nsf, and domcfg.nsf

databases.

Ironically, while Lopht was posting the advisory about

the security-flaw on its website, Lotus outlined details

of its E-commerce and public-key infrastructure

security plans at its developer’s conference. To address

security concerns, Lotus placed its IETF PKI compli-

ant implementation in the public domain last summer.

Lotus officials said Microsoft, Intel, and Security

Dynamics Technologies pledged to support the

reference implementation. Having a single PKI imple-

mentation will keep down the number of certificate

authorities businesses will have to maintain for trading

partners and foster greater extranet development. PKI

functionality for Notes/Domino will be completed

and available in the next six to 18 months.

European Companies Not Impressed With E-Commerce

A recent survey by Andersen Consulting shows that many European business executives are slow to incor- porate electronic commerce into their operations. While 82% of executives said they believe E-com- merce will have a strategic impact on their business in the future, only 39% are taking steps today to incor- porate the technology into their strategy. Only 19%, the survey found, regard E-commerce as a serious competitive threat.

Moreover, one-half of the respondents believe that

consumers lack an understanding of E-commerce. A

majority view privacy and security as major barriers.

Eight percent cited the need for governments to work

together to form a common, international E-com-

merce framework. The survey, conducted between

December 1997 and July 1998, involved more than

300 senior executives throughout Europe

Vendor-Supplied Security for Computers Dell Computer Corp. is working overtime to help IS

managers secure the data in their users’ notebooks,

desktops and servers. Dell recently announced

DellGuard, a security initiative that features password-

protected hard drives for notebooks and desktops,

along with an 800-number that assists in tracking

stolen PCs. In the first half of 1999, the company plans

to add smartcard hardware to its notebooks and

desktops that will provide a single point of user

authentication.

The smartcard solution, which would require a card

reader and be used during logon, would reduce the

number of passwords needed and lessen the chance of

an unauthorized user gaining access to data stored on

a PC or network. Dell’s smartcard security product is

expected on Latitude notebooks and OptiPlex

desktops, sources said. While the company is consider- ing using third-party smartcard devices, it is also devel-

oping its own reader, which could add about $100 to

the price of the notebook, sources said. Dell is also

evaluating biometric security technology, including a

fingerprint sensor developed by Veridicom Inc., in

Santa Clara, Calif., which authenticates users via their

fingerprints.

The company’s DellGuard initiative will focus on

standards-based solutions that are compatible with Microsoft Corp.‘s Windows NT 5.0. Windows NT

5.0, for example, will support smartcards and include a smartcard API as well as built-in data encryption.

Dell isn’t alone in its security efforts. Compaq

Computer Corp. is a member of the BioAPI Consortium, a group that is working to develop APIs for biometric security. The Houston company also

566

Page 4: Windows-98, RIP and my Adrenaline Rush

Computers & &writ-v, Vol. 17, No. 7

has a fingerprint recognition device for its Deskpro

PCS.

IBM will also introduce a smartcard encryption prod-

uct for its ThinkPad notebooks.The company is devel-

oping a reader that will tit into the notebooks’ PC Card readers. The smartcard product, which has not

yet been priced, will be available first as an option,

sources said.

Hewlett-Packard Co., a longtime smartcard backer,

recently announced ProtectTools, a program to inte-

grate smartcard readers into its Vectra VL desktops,

Kayak workstations and OmniBook notebooks. The

Palo Alto, Calif., company is working withveridicom’s

sensors for inclusion in future desktops and note-

books.

While all of these initiatives are useful and beacon the user to thinking more about being secure and less

about the overall problem (it’s the “I have a lock on

the door and am therefore secure” syndrome), they are

all based on different methods, technologies, tech-

niques, etc. Until we can get all the vendors together

in a single mode of security, all this does is exacerbate the differentiation of security products for the same

technical problem and, as usually happens, the users

turn the security features off in defiance of IS person-

nel due to their “operational needs” (i.e. they don’t

want to mess with the solution as it is intrusive and

annoying).

Network Associates Introduces a New Firewall Concept: Adaptive Proxies

Network Associates, Inc., recently announced a new method to protect networks via firewall technology

that promises to remove the longstanding tradeoff between security and speed when choosing a tirewall.

The new patent-pending tirewall technology features

‘Adaptive Proxies’ which maintain the tight security

standards of proxy firewalls, but can dynamically ‘adapt’ packet flows on the fly to achieve substantial

performance improvements. The new Adaptive Proxy technology is the result of years of research at NAI Labs, the security research division of Network

Associates.The new technology will debut this month

in Gauntlet Firewall 3.0 for Windows NT, and will be

widely available on all Unix and NT versions of

Gauntlet later in 1998.

Historically companies have faced a tradeoff between the speed of stateful packet inspection firewalls and the

tight security of application proxy firewalls. The new

adaptive proxy model eliminates that tradeoff by

dynamically applying the appropriate degree of secu-

rity as it is needed.

Application proxy firewalls like Gauntlet have tradi-

tionally been viewed as better than average security

due to specific architectural features that help secure ancillary information when a connection is in

progress. Because all data passing through the firewall

is examined at the application layer, the highest level of the protocol stack, application proxy tirewalls have full knowledge of exactly what is occurring in each

attempted connection. Proxy firewalls are also consid-

ered to be superior to stateful packet inspection

firewalls because they act as a ‘proxy’ for all authorized

connections, never allowing direct contact between

the trusted and untrusted networks. Although these methods offer significantly tighter security, the addi-

tional security measures sometimes require more time

to process.

In contrast, stateful packet inspection firewalls to

simulate the approach of an application proxy tirewall

by examining data through a proprietary inspection

module at a much lower level of the protocol stack. Once a connection has been established, stateful pack-

et inspection tirewalls also allow direct connections (if

NAT facilities are not enabled) between endpoints

through the firewall, potentially exposing internal systems to compromise from sophisticated attackers if the firewall is incorrectly set up.

The new patent-pending ‘Adaptive Proxy’ technology

supposedly gives users the best of both previous fire-

wall technologies. Adaptive Proxy firewalls dynamical- ly ‘adapt’ packet flow on-the-fly based on user-defined security rules. This allows users to customize firewall policies to their specific needs without sacrificing speed or security. When security requirements are

567

Page 5: Windows-98, RIP and my Adrenaline Rush

Security Vie ws/Dr. Bill Hancock

high, the initial security examination still occurs at the

application layer, assuring the maximum security of a

traditional proxy firewall. Once all the details of that

session have been cleared by the proxy, however,

subsequent data packets can proceed directly through

the much faster network layer. Initial vendor bench-

mark tests of the new Adaptive Proxy technology have

demonstrated tenfold or greater performance

improvements with zero security compromise (of

course, these are vendor provided and still need to be

verified).

Gauntlet’s new Adaptive Proxy technology will also

enable more flexible integration between individual security products such as security vulnerability scanners, virus security scanners, and intrusion protec-

tion sensors. As part of its ‘Active Security’ initiative,

Network Associates is enabling properly authenticated

devices to automatically ‘adapt’ firewall security levels

according to a firewall administrator’s pre-established

security policies whenever security sensors and

scanners detect an important threat to the network.

Next!

Aussies Outsource their Equivalent of the US National Security Agency

The partial handover of Australia’s most secretive intelligence network to a foreign company was a

threat to national security, the main opposition Labour

Party said Tuesday.

The Defense Signals Directorate (the Australian

equivalent of the US National Security Agency),

which is so secretive its annual budget isn’t even made public, has handed over some of its foreign intelligence

gathering to private companies, one of which is

British Aerospace Australia. The handover was made public when British Aerospace advertised for 40 peo- ple with expertise in languages of the Asia-Pacific region to work in Australia.

Labour defence spokesman Arch Bevis said handing

such sensitive work to a foreign-owned company was madness and clearly not in Australia’s interests.

“Not even Margaret Thatcher or Ronald Reagan

went this far with privatization”, Bevis said. “Giving

British Aerospace Australia a contract to advertise for

and employ language and communications intercept

experts has enormous implications for Australia’s

national security.”

“In a conflict situation, the information they would be

gathering could affect the lives of thousands of military

personnel.” Bevis called on Prime Minister John

Howard’s government to make public any plans for the

sell-off of other aspects of Australia’s defence systems.

What I find truly interesting about this is that the task

is over and done with, the politicians are going crazy over it and there is a high probability that the situation

has existed for years. In my years working in ‘spook

shops’, it was not uncommon at all to have a great many commercial contractors working with us side-

by-side. It would never appear on the books of any

contractor as to what, exactly, they were doing - but they were there doing it nonetheless. Sorry, guys -

nothing really earthshattering here. Been happening

for years. I sometimes wonder what will be outsourced

next. Upon reflection, I probably don’t want to know.

New E-mail Newsletter on Electronic Identify Fraud

John Ellingson, principal of e-Dent&cation LLC

announced the start of publication of a free E-mail

newsletter called Electronic Identity Fraud.

Publication is devoted to the newly created problems of electronic commerce and electronic data inter-

change (EDI) wherein people in remote locations

do business through electronic means and lose the

valuable opportunity for personal face to face evalua-

tion of each other.

Much attention has been devoted to the security of the message, using encryption and passwords and fire- walls. But the important question of the identity and honesty of the sender of the message has been

ignored. The newsletter is available on request to [email protected]. Please say Identity Fraud as subject of request.

568

Page 6: Windows-98, RIP and my Adrenaline Rush

Computers & Security, Vol. 17, No. 7

John Ellingson is chairman of NBIB Inc. which devel-

oped an identity detection system now used by 17 000

banks.The new system is designed specifically for non

bank companies engaged in electronic commerce. He

is at [email protected].

World’s Smallest Combination Lock is Created

A team of US scientists have developed a minuscule

mechanical device they describe as “the world’s small-

est combination lock”, promising to build a virtually

impenetrable computer firewall.

Sandia National Laboratories said in a news release

that the Recodable Locking Device is a series of tiny

notched gears that move to the unlocked position

only when the right code is entered.Using the micro-

electromechanical system (MEMS) so small that it

takes a microscope to see it, Sandia said, the device is

the first known mechanical hardware designed to keep

unwanted guests from breaking codes and illegally

entering computer and other secure systems.

The Recodable Locking Device is hardware. With it

in place, a user would only have one opportunity to

enter the correct password - and a one in one

million chances of guessing it, compared with a one in

10 000 chances in most passwords used in software

firewalls. The system shuts down if the password is

incorrect and can only be reset by the owner. The

entire device is about the size of a button on a dress

shirt and could be built into a small chip that would

be incorporated into any computer, computer

network or security system.

It consists of a series of six code wheels, each less than

300 microns in diameter, driven by electrostatic comb

drives that turn electrical impulses into mechanical

motion.To unlock the device, a user must enter a code

that identically matches the code stored mechanically

in the six code wheels. If the user makes even one wrong entry, the device mechanically ‘locks up’ and does not allow any further tries until the owner resets it. Sandia said that the device has a powerful potential besides being a deterrent to hackers. They expect to

see the device used in commercial applications within

the next two years.

Getting Bitten by Year 2000 Problems in Places You Never Thought of.. .

Oddly enough I get a lot of E-mail questions about

security and the Year 2000 problem. While some of

you readers will refer to this section as “I already knew

that”, and you may be correct, it strikes me that there

are many out there in cyberspace who really don’t

understand the Y2K problem to the depth that it

extends. If nothing else, use this section of this month’s

column to educate your management.

Everyplace you read you hear about the Year 2000

problem, especially when computer systems and

software is involved. Is it real? Yeah, unfortunately, it

most certainly is. Is your company at risk due to it!

Yes, it is and that’s a question we can explore in detail

later, but first we need to understand what the Year

2000 (also known asY2K) problem is.

When programmers write programs, which are on

every computer in every location, they have to create

program code sequences that involve the use of dates

and times. When defining any type of date many

times, programmers must specify rules for the format-

ting of year of the date. For instance, specifying only

19 for a year as the first two digits is going to cause big

problems when the year 2000 hits and it starts with

20. Programs cannot do anything other than what

they were programmed to do and those who are

expecting 19 for the first two year digits are going to

freak out in a major way when something hands a 20

to them instead. At a minimum, things stop working. At a maximum, systems crash and incorrect date

information is saved. This can have some pretty far-

reaching and catastrophic effects for a lot of things like

databases and real-time programs which use dates and

times for serious work like nuclear reactor failsafe

programs.

Even worse are where programs have algorithms where date computations are made based upon only the last two digits of the year. Let’s say you wanted to

569

Page 7: Windows-98, RIP and my Adrenaline Rush

Security Views/O=. Bill Hancock

know how many months had transpired from the first

of January of the year 1900 until the last day of

December of the year 1999. If the computation were

made such that only the last two digits were used, then

this task done before the year 2000 would result in a

computation of 99 times 12 or 1188 months. But,

what if the same question were modified to the

number of months from the first day ofJanuary of the

year 1900 until last day of December of the year 2000?

The answer would come back as 12, not 1200 as the

correct answer would be.This simplistic example can be expanded to problems like invoicing for time spent

on jobs (typical timesheet applications), time manage-

ment software, calendar tracking applications, etc.,

which affect everyday user routines. It can also affect

railroad scheduling, flight scheduling, reservations for

practically anything, industrial control systems, finan-

cial and accounting systems and, basically, anything

run by a computer.

So, the problem is quite serious. And, without updat-

ing systems, it is not going away. In fact, the two

examples above are simplistic ones that are easily

understood.There are a rash of them that are consid-

erably more complex and difficult to understand and

correct. Another problem are the nay-sayers that claim

that Y2K is over-hyped and that very few things will

be affected.There have been some articles written by

non-technophiles to this effect and they couldn’t be

more wrong if they tried. There are extremists that

claim “all will come to a halt” and those who claim

“nothing is wrong”. The truth is in the middle,

leaning more towards the “all will come to a halt” than

“nothing is wrong”.

Some companies are taking the Y2K problem very

seriously. Consider the expenditures of the following

banks to address the Y2K problem (sources are

Chicago Tribune and Grain’s Chicago Business):

Chase Manhattan Corp $250 million Citibank 50-70% of staff dedicated

to the problem BankAmerica Corp $250 million First Chicago NBD Corp $100 million LaSalle National Bank $30 million

If you examine the locations where computers are

being used, you can begin to see how problems with

something as simple as a date can cause personal and

professional problems.

I have an intercom system in my home. It has a digi-

tal calendar in it and was manufactured about two

years before I bought my home in 1987. Just for fun,

I moved it up to January 3, 2000, and it displayed

January 3, 1900. It does not even have the year 2000

capability to display the number in the unit. So, I

called the company that makes the unit and they told

me that to fix the problem would require replacement

of the main unit with a new unit for a cost of about

$700. Needless to say, I am really unhappy about this

and am now tracking down the programmer for my

intercom unit to do him/her bodily harm.

Another personal experience is my microwave oven in

my kitchen. Installed at the same time, it has a timer,

date and day display. Guess what happens when you

set it to January 3,2000? It beeps at you and displays

a fault code and then reverts to 1900. I called the manufacturer and they said that a $50 upgrade will fix

it. Gee, how happy am I again? I am up to $750 in

personal out-of-pocket upgrades and I have not even

gotten to the house security system, car calendars,

sprinkler system,VCR clock/calendars (yes, campers,

they are affected as well), refrigerator (yes, it is ‘calen-

dar-challenged’), stereo system, digital wrist watches,

electronic alarm clocks, etc. All of this has nothing to

do with my office, but all of it is up-close and person-

al. And, a lot of it is going to break in the year 2000. Sigh.

What does this have to do with security? Plenty.

Remember that most computer-based security tech-

nologies use programmable date information to keep

logs and audit trails. Let’s examine some very simpli- tied example areas of risk and security issues where the Y2K problem, not fixed, will cause problems:

Perimeter facility access security products. These products, typically used to secure a building or grounds area, use computers to control access to the facilities. Dates are extremely important for enforce- ment ofTime of Day (TOD) operations when people

570

Page 8: Windows-98, RIP and my Adrenaline Rush

Computers & Security, Vol. 17, No. 7

are allowed in or out of a facility. TOD also causes

problems for people with access which expires on

specific dates at specific times. If the date modules of

the programs controlling perimeter access can’t deal

with Y2K, the entire security system may be in jeop-

ardy and will either lock everything down or lock up

nothing at all.

Card control access systems. In a recent test at a

customer site, we put the clock ahead to test a card

control system. Everything worked great except the

elevator card key facilities. Turns out that the system

was upgraded properly, but the elevators were not and

it was expensive to do so. Therefore, it was cut from

the budget. Since the card keys did not cooperate in

the elevator card reader PROMS after Jan. 1,2000, the

elevators would not allow us to go anywhere a card

was required to reach a specific floor. I am sure that

other embedded logic card control systems will have

similar problems as I have seen them on several

already.

Fax machines. Many of them have built-in electronic

day and date display and transmittal facilities.We took

a relatively new unit (about 3 years old) and moved

the date up just for fun. It wouldn’t let us and beeped

at us every time we tried. Really frustrating. In many

companies, the date-time-group (DTG) provided on

transmitted .fax documents is a crucial business

requirement, especially where negotiations are going

on or where there are date restrictions on actions by

company management. If a fax machine cannot trans-

mit an effective date, this is bad science for all

involved.

PBX systems are particularly sensitive to calendar date

events. Discussions with some industry experts tell me that many PBX switches that are over 10 years old will

have problems withY2K. A customer of mine running

a very large switch with over 10 000 users has a switch

that is effectively 20 years old. Needless to say, aY2K

test was a disaster and upgrading the switch will be

expensive and time consuming.

Manufacturing systems. Companies that engage in manufacturing products actually have three threats to their business: embedded systems, the supply chain to

the business (and all associated systems) and internal,

third-party packages. There are a lot of 1970’s tech-

nologies in embedded systems that will not work

properly in 2000 and there is a great deal of microcode

and other specialized applications that may be difhcuit

if not impossible to debug and fix. In fact, one well

known pharmaceutical firm I know of is discontinu-

ing a version of a major diabetic therapy product due

to the expense of converting the system to work in

Y2K: it’s cheaper to get rid of the system and move

customers to a newer product than to upgrade the

system which makes the pharmacological components

to work properly later on. Supply chain vendors who have problems shipping raw materials will impact

arrival and scheduling of manufactured components

and cause total chaos in the manufacturing and deliv-

ery of product. Internal packages configured for Just

In Time (JIT) inventory systems may suffer due to the

standardY2K problem but also may not be equipped to deal with the issues of scheduling and manufactur-

ing problems imposed by materials problems with the

supply chain issues. It becomes a mess rather quickly

when just a few things start to go badly.

Retail systems. Solely dependent upon supply chains,

the problems of supply chain system failure to the

retail channel will have far reaching and major effects

to retail systems which depend on rapid delivery of sellable items where inventory is tight and rapid

change in spending patterns occur. Supply chain

components may not notify the retail consumer until

it is much too late to do anything reasonable about the

problem.

Utility systems.What happens if a nuclear power plant

failsafe system, which may not be date-dependent, is

told by a date-dependent system to scram the reactor. At that point, massive logging of events is required and

needed. What if the logging system fails? The plant may be left in an ‘Unanalyzed Condition’ (which is

very bad - all safety issues that happen at a nuclear

plant must be completely analyzed and resolved before

a restart can occur) and the plant must be shut down until the safety issues are identified and corrected. Nothing has caused melt down and probably won’t. The problem is that it cannot be started back up. Lights out.

571

Page 9: Windows-98, RIP and my Adrenaline Rush

Security Vie ws/Dr. Bill Hancock

Health care. Hospitals and other medical organizations

depend on diagnostic systems which have their own

embedded systems within them. A lot of these systems

stop when maintenance intervals are reached.

Computing the wrong date may cause that to happen

a lot quicker than expected.These organizations have

a lot of the same problems as the retail sector’s

problem with supply chain Y2K problems, but here

people’s health and, indeed, their lives may be at stake.

The list of opportunities to fail goes on and on.The

point is clear: Y2K is serious, ubiquitous and some-

thing that a lot of companies are simply not paying

attention to solving.This is especially the case in small

and medium sized companies where expenditure to

fix the problems is not part of the overall corporate

goal of survival.

Then, there are the legal problemsYes, legal problems.

In any area where liability is produced, there is always

the human and corporate tendency to find someone

else upon which to shift the blame and, therefore, the risk. For instance, in a public company, failure to

disclose Y2K potential problems subjects the directors

and officers of a company to a rash of lawsuits, The liability is that these individuals have a fiduciary

responsibility to act in the best interests of the corpo-

ration. While corporate ‘standards of care’ vary from

state to state, they exist to protect the company and

stockholders. What is particularly important, especial-

ly for public companies, is the risk to the company of

shareholder lawsuits and enforcement actions by state

or federal authorities based upon a company’s lack of

official or sufficient disclosure of Y2K issues in

required public filings with the Securities and

Exchange Commission (SEC), or with state securities regulators. The SEC issued guidance in 1997 (Staff

legal Bulletin No. 5 onYear 2000 Disclosures) which

advises companies of their year 2000 disclosure obligations. While a ‘guidance’ by the SEC is not a law,

you don’t ever ignore it lest your company trigger an SEC enforcement action or class action lawsuit by stockholders. The SEC requires “specific and mean- ingful”information aboutY2K issues and also specifies minimum informational requirements (which can be extensive in some situations). By keeping a consistent

and exhaustive chronology of events that the compa-

ny undertakes to properly address the problems, these

opportunities are minimized. Lack of doing anything

can open up the company to a variety of problems

including business failure.

Ok, now that your consciousness has been properly

raised, let’s examine what you can do about the prob-

lem and what steps are necessary to avoid the Y2K

pitfalls that are sure to come along.

There are four overall steps that have to happen to

properly address and fixY2K problems:

1. Inventory and assessment of your exposures toY2K

in software and systems .

2. Analyze and find your year 2000 risks and legal

requirements.

3. Fix your programs and applications.

4. Test your changes.

Before you get too carried away and start the steps,

there is that legal liability exposure-‘thing’ that you

have to be concerned about - whether you are a

public or private firm. One sure help in a courtroom

is a great deal of documentation about all the steps and

work that was done to ensure that your systems were

being corrected for Y2K compliance. That starts first

and now: document everything that goes on, regard-

less of how trivial it might seem, to ensure that your

company has proper paper-trail compliance efforts in

the case of a legal action against the company as a

whole or specific officers and directors.

While the following statement might seem a little brain-damaged and normally filed in the ‘common

sense’ file, it’s crucial: upgrade and fix mission critical systems first! A quick assessment will yield information about what systems and software are critical to keep- ing the business rumling or what supply chain facili-

ties are critical to keeping materials flowing. These areas obviously must be the first addressed regardless of

572

Page 10: Windows-98, RIP and my Adrenaline Rush

Computers & Security, Vol. 17, No. 7

their complexity as they are considered to be business

critical to keeping the company afloat as a profit gen-

erating machine.

(1) Inventory and assessment of your expo- sures to Y2K

This step has two components: business issues and

technical issues. Business issues require the concerns

over compliance with Y2K by vendors, business part-

ners, suppliers, subsidiaries, embedded systems, end

products, retail products or provision and industry-

specific business, legal issues and regulatory require-

ments. The technical issues include: an inventory of

software products in-use; vendor statements verifying

either no issues or what issues are to be solved inY2K

compliance (you will need to contact each one); in-

house software analysis to discover what was ‘home grown’ and will cause problems; embedded systems

that you may be using that have older software in

them and may fail or not operate correctly; etc.

(2) Analyze and find your year 2000 risks and legal requirements

Now that you have a rough idea of what the scope of

software and business problems are to be solved, the

problem of analyzing any in-house code and process-

es for Y2K conversion efforts needs to be done. This

can be an especially painstaking effort and is essential

to ensure that you know exactly how much work and

what types of efforts will be required to properly con-

vert systems to Y2K compliance. Testing of date

changes and effects of failure of one system vs. what

happens to another is just a small piece of what has to

happen. Other tasks include how your company

affects supply chains, your requirements onY2K legal-

ly and to your customer base and many other related

issues.

(3) Fix your programs and applications

This sounds easy, but it’s not. A lot of programs that are home-grown or are used in systems no longer sup- ported by a vendor may be extremely difficult, if not impossible, to fix. This may entail entire replacement of selected systems or subsystems that are in use to

ensure that they can properly function when year

2000 comes around.You will most likely need to find

source code for affected programs or get the upgrad-

ed ones from vendors if they exist. Consultants who

are familiar with your systems and problems will most

likely need to be retained and an overall project and

plan for correction will need to be designed and

implemented.Tools and other facilities that will facil-

itate the conversion of code or products will also need

to be tested. certified and used for the conversion.

(4) Test your changes

This is often one of the most painful stages and often

takes as long as it takes (or more) than it does to make

code changes. Testing involves the use of automated

testing tools., conditional testing, interrelationships

with other programs and how the changes affect other code components and many other issues. If you are in

the supply chain as a vendor, the pain can increase

seriously when you consider that you may need to

create a parallel system environment to test changes

and updates. Interactive systems can be a real challenge

as user interfaces, reports, database interfaces and all

manner of interactive methods must be tested to

ensure that everything works as required. Of course,

there is user training, documentation and many other

steps to ensure that everything gets done correctly and

functions in accordance with plans.

A final comment about consultants. Be careful who

you select to do your work, ensure they have the prop- er credentials and get some specific information about

what they can and cannot do for you in the conver-

sion effort. As in any service sector industry, there are

some very reputable individuals and some that are less

than what they seem. Remember that all the work is

being done by people, not machines, and the selection

of those people with the right qualifications are what

makes or breaks a conversion. Also, the customer has

the absolute right to know what they are getting and

why and also be intimately involved in the process.

Failure to get thoroughly engaged in the process leads to miscommunications, overcharges and runaway pro-

jects. Get involved and stay that way.

When starting your conversion efforts, ensure that

573

Page 11: Windows-98, RIP and my Adrenaline Rush

Security Vie ws/Dr. Bill Hancock

your team has the proper systems, tools and facilities

that are necessary for the job. Some larger consultan-

ties have developed their ownY2K conversion pack-

ages. Mainframe vendors, such as IBM, have entire

Web sites and conversion suites that are used for

conversion help. Project management tools, Gantt

chart progress tracking, source control systems, formu-

la and strategies for date management and upgrades

and a whole host of other technical tools and facilities

are necessary to properly plan and upgrade systems

and code forY2K compliance. All of this costs money

and does not come for free. These are expenditures

over and above the actual code conversion and, with-

out them, the effort will take much longer and have a

less overall chance for success.

Y2K is non-trivial. In some systems, the effort may

involve as little as upgrading a system to a new version

of the product. In most cases, especially if there is

home-grown software involved, the effort must be

carefully planned and progress controlled to ensure

success. Even if your in-house systems are not affect-

ed, if you are a manufacturer, retail supplier or other

‘middleware’ type of business, you may be affected by

otherY2K problems at other companies and vendors.

Take some time and analyze your exposures and know

what alternatives you have before real problems creep

up on you and it’s too late to do anything about them.

Insist on vendor compliance and remember to work

with your vendors and suppliers to minimize the

impact of systems that have not been converted.

Develop a disaster plan of action in case there are

supply-side problems that you have no control over

but affect your ability to pursue your business. And,

remember that there are a great deal of legal issues that

you must deal with to ensure that the company’s liabilities are properly dealt with.

Recommended reading

The Year 2000 Software Crisis Ian S. Hayes, William M. Ulrich Yourdon Press Computing Series ISBN O-13-9601 54-6

Practical Methods for Your Year 2000 Problem

Robert B. Chapman

Manning Publications Co.

ISBN O-884777-52-X

The Year 2000 Computing Crisis

Jerome T. Murray, Marylyn J. Murray

McGraw-Hill

ISBN o-07-912945-5

Dr. Bill Hancock, Executive Vice President and Chief Technolog Officer of Network-l Software and Technology, Inc., is a well known computer and network consultant, designer and engineer with thousands of network desqqx to his credit. In the business for over 25 years, he has drslgxd and rr-engineered networks (over 4000) for many of the Fortune 1000 as well as many international companies and governments with system counts from two to over 1.5 million rystrms. He has held full-time technical and management positions at various Fortune 100 companies including Standard Oil of Ohio, I>igital Equipment Corporation, Texas Instruments and US governmental organizations such as the Naval Security Group Command. A prolific network architect and designer, he has desiqed networks for a wide variety of organizations such as the Capitol of the United States of America. 17 power companies, NASA research networks, aircraft control systems such as components of Boeing aircraft and the F-16 and F-22, manufacturmg networks, K&II networks, telephone companies, banks and financial institutions, distributed control systems, various governmental networks and components of the worldwide network known as the Internet. A network and system security expert, Bill has designed and developed commercial dial-up security, encryption, network firewall, authentication, digital signature and other products.As a consultant, Bill is often sought to provide guidance on security policies, procedures, trchnolo+s, strategies and actual hacker prosecutions and trackdowns. Bill often works with law enforcement professionals worldwide to identify, stop and prosecute computer criminals and offenders. Bill is an often sought speaker for keynotes at InterOR Comdex, CEBIT, NT World, NrtworksExpo, Compsrc, Internet World, Mactivity and is well known for his detailed knowledge of networking and security as well a\ his humorous style of speaking. Uill has written 20 books on computer networking and security and has wrlttrn art&s for Datn C(~rwrrr~rnirati[l,r~s Mqaziue, DEC hf&ionnl, D@fa/ I\%~Lx, ?+lru 34/3X, Tl~e Wall Street J~umnl, 7%~ D&s .IL&ri<q X&s, IEEE ~l’etuarkj, ~X’c’ettrark M/;,rld, .%eruvrk Set&y and many othrr publications. Hr currently writes a regular column in h’etu~or~ Sctlrrity magazine. Hr is also a US network expert to the IS0 and sits on various international ctandards committrea. 13111 is a member of many industry societies (IEEE, ACM, I)ECUS, etc.) and has sat on the boards of several organizations. Bill is a member ofANS1 and sits on srvrral standards committees domestically and internationally, Hr holds \cveral patents in networking .md security trchnologics and is a Certified Informatmn Systems Srcurrty Professional (CISSP), Certified Network IIrsi~mnrr (with Archltrct Endorsement) and has earned a B.A.. M.S. and I’h.lI m Computer Science. Further bmgraphical information can br found in: W/IO? W/IO it1 rlrc K&/d, K%o i M/l10 i,r Awleritn, l&%0’, l&%0 irr Srir~lrc 0ifn Eqiwerir~~~, otrd W7raIc W/r0 ifi Firmrrr md Irrdurtry.

574