Slide 1

Michael Mainelli, The Z/Yen Group

[A version of this article originally appeared as "Defence and the Internet", Oman Telecom Book ’97, (Proceedings of Oman Telecom ’97 Conference, September 1996), pages 73-78, Trifoil Publications Inc (January, 1997)]

DERA leads much of the UK defence and security work on the Internet and other computer-based systems and networks.  Some of the highlights in DERA's Internet history include being the third international site for the Internet (Arpanet) outside the USA in 1978 and the first Internet site in space (STRV) in 1995.  DERA's computing record ranges wide and deep, from the invention and continuing development of liquid crystal displays through optoelectronics and light-emitting silicon on to advanced operating systems and languages.

The Internet is a child of the defence community, so it is little surprise that it remains an important technology area for defence research.  What has surprised many people has been the explosion in interest among the commercial community.  Estimates of users in the USA alone run from 24 million to 37 million, over 90% of them connected only in the last five years, and few of them from the defence community.  Estimates of worldwide users exceed 50 million.

At a technical level, the Arpanet of the 1970's is little different from the Internet of the 1990's.  However, a number of factors have changed to encourage the explosion of usage:

1970’s

1990’s

·         limited community of defence and academia, cost/benefit of connectivity not high

·         millions of users and perception of wide community outside defence and academia

·         separate telecoms and computing capabilities

·         integrated computing and telecoms

·         e-mail and limited file transfer applications

·         additional end-user applications, e.g.  browsers, USENETs

·         off-the-shelf tools and connectivity

·         specialist market

·         large corporate market

·         enormous potential consumer market

·         relatively high cost

·         largely sunk costs for infrastructure based around PCs

 

Internet and Today's Defence Community

DERA is on the Internet every minute.  At the most basic level, DERA researchers use the Internet as part of the international scientific and technological community.  As researchers world-wide use the Internet to broadcast and receive commentary on developments, so do DERA's scientific researchers.  DERA broadcasts its capabilities to the wider community through a World Wide Web site (http://www.dra.hmg.gb/).  DERA provides advisory services to other UK government departments on their Internet connections, not just security aspects, but also network capacity, telecommunications purchasing, application suitability and inter-connectivity to other networks.  DERA's advisory work includes other governments interested in ensuring that Internet usage is controlled, e.g.  firewalls, domain policing, usage monitoring or material content control.  DERA advises commercial concerns such as large telecommunications service providers, financial institutions, local governments, tourist agencies and a wide variety of other users, especially where security, robustness and control are important.  DERA has even moved to providing some managed services for users with special requirements.

DERA has piloted Internet concepts such as improving internal communications by setting up an internal Internet within the organisation, a so-called Intranet.  DERA's Intranet connects 5,000 internal researchers on a secure network using browsers and usenet/newsgroup software from off-the-shelf commercial packages.  The Intranet permits users to network with each other in multi-disciplinary activities to bring to bear the full power of a large organisation.  DERA's configuration of commercial packages for internal use is novel and has proved robust in practice.  Some estimates of cost/benefit imply that some of the greatest rewards from Internet systems for DERA may come from the Intranet.  DERA is working hard on the integration of multiple information sources, combining Internet, Intranet, CD-ROM and corporate databases.  An open source information product is available, DELOS, which uses hypermedia display, sophisticated search and retrieval, user interest profiles and automated information mining techniques through secure connections.

The Internet has not ceased to be a defence tool.  DERA research covers a variety of subjects which focus or touch on the Internet.  A small selection includes: 

  • Security - e-mail, firewalls (e.g.  Heimdall), penetration testing, anonymous use, mirroring/caching, encoding algorithms; 

  • Services - mail servers, search engines (e.g.  DELOS), browsers, agents, brokers, newsfeeds, data compression, service integrators; 

  • Structure - system architectures, design/build/test methodologies, web page design standards, information perception; 

  • Standards - object-oriented standards, telecommunications, compression, network protocols.

In contrast to being primarily a provider of basic research a few years ago, DERA today spends much more of its time ensuring that it is up-to-date with developments as commercial applications burgeon.  In a similar vein, DERA is much more interested in the availability of defence solutions through purchasing "commercial off the shelf tools" (COTS) to build applications rather than developing in-house.  A good example of this has been DERA's development of the Intranet using standard packages as described above.  COTS is a trend in a number of defence procurement areas.  As the development time of new information technology applications shortens from years to months, the military has to reduce its cycle time for development and its acceptance time for external applications safely if it is to achieve benefits similar to those that commercial organisations hope to achieve from the use of Internet applications.

Open networks and their evolution have direct military and economic importance to nations.  Several of DERA's operational analysis studies, and those of other military institutions, have highlighted the potential vulnerability of networks to external penetration or even external closure, the so-called "Info-War Scenario".  Critical network vulnerabilities include not just government networks but also those used in the financial services industry, utilities or travel.  In the extreme, e.g., no hospital records, transport, etc., disruption of all aspects of our digital society would bring us back to a primitive state.  While the Info-War threat justifies many offensive and defensive areas of research, it also requires closer contact with the civil telecommunications and financial institutions who are most likely to be affected.  As just one example, in order to promote closer civil contact, DERA initiated a Financial Laboratory to work with financial institutions in the City of London on financial risk management, one area of which was `hardening' of systems to external threats.

Future Implications

The Internet changes and expands so rapidly that one declines to predict what may occur.  Nevertheless, some areas of research indicate possible directions in which technology may move.  Robert Cringely describes "killer applications", in his book Accidental Empires, as those applications which turn an interesting technology into an explosion, e.g.  VisiCalc for Apple, word-processing for many systems, and Netscape for the Internet.  Predicting killer applications is hazardous but thought-provoking.  None of the following descriptions of potential killer applications has any official sanction whatsoever.  The descriptions are put forward to indicate some of the after hours musings which do occur.

E-cash (various forms of electronically based media of exchange) is one such potential killer application.  Much has been written about e-cash, and many pilots are in progress, e.g.  Mondex in the UK.  The ability to combine commercial exchange with Internet information and services will be powerful.  The technical feasibility of wholly digital safety and control remains in question.  As powerful as encryption techniques or public key algorithms become, they remain vulnerable to subversion through wholly digital means.  At present, acceptable secure systems require non-digital elements working in conjunction with digital security, e.g.  signatures on credit cards.  Further concerns surround the legal and enforcement issues.  At the same time, a successful e-cash application could, almost at a stroke, develop a fully-fledged currency overnight with a market size which might overtake minor currencies in the international markets.  With true e-cash, i.e.  a store of value, a trusted medium and convertibility, a petrochemical company could easily issue petro-backed vouchers worldwide to fund its business.  E-cash could fundamentally change the nature of international commerce and reduce the ability of governments to control their own currencies.

Another potential killer application is the development of end-user translation software enabling web pages to be viewed in a desired language, e.g.  English web sites viewed in Arabic with the translation performed dynamically by computer.  Much early research on natural language translation is being revisited as the previous obstacles, i.e.  a lack of digitised material and a critical mass of international readers, recede.  With a world market, opportunities for natural language translation are enormous.  A few basic translation products already exist.  End-user translation software will make it more difficult for governments to control information received by their Internet-literate citizens.

"We want (secure) bandwidth on demand", has been the cry of large corporate network users for several years.  Large corporate users do not want to build large, expensive international networks where base size is calculated on handling their largest fluctuation in demand.  They want to buy network capacity as required.  A related, but different, cry has been for "distance independent prices" analogous to charging regimes for intra-national post in most countries.  The Economist recently devoted a telecommunications survey to the effects that distance independent prices might make to economies.  These two cries, if satisfied, could lead to a radical change in the way defence networks are bought.  Just as for corporate organisations, defence organisations need access to large bandwidths at a distance, but not all the time.  Multi-cast technologies, permitting complex combinations of broadcast, unicast and multiple bandwidth interactions, are likely to be at the forefront of technical capability and price reductions in networking.  The importance of pricing in Internet development can be seen in the high-sensitivity of Internet usage to local call pricing in the USA.

At the same time as networking costs reduce for everyone internationally, the ability of terrorists, pornographers, criminals or other undesirables to construct world-wide networks also becomes easier - food for thought as both sides of a conflict increasingly use the same infrastructure.  The widespread availability of long-key cryptographic systems on the Internet, which will eventually form the basis of secure commerce, currently protects some communities with less civic goals.  The intriguing problem for civil and military guardians who desire control is to intercept and decode dangerous material while not compromising the intrinsic security of commercial systems.  A related application area under development is the `glass wall'.  Glass walls, if successful, may permit nation states to control net access both in and out, thus allowing some cultural barriers to be erected.  In order to be effective, glass walls are likely to require co-ordination with telecommunications tariffs and local service levels.

Much research has been devoted to `agents' or `brokers', independent software programs acting through the Internet for their `client' to find services.  Examples of such agents abound, booking air travel, finding lost friends, looking for interesting news, recommendations for new books or music suitable to the client, etc.  The ability to integrate internal and external information through search engines (e.g.  DERA's DELOS system) has led to so much information being available that active submission to users must replace passive perusal by users.  Tentative attempts to develop automated compression or precis/digestion of information, e.g.  personal newspapers, have already had some limited success.  Seeking services through the Internet is becoming standardised through standardisation of brokers, e.g.  CORBA (Common Object Request Broker Architecture).  Brokered services could grow rapidly and profitably, e.g.  housing searches, to property databases, to legal advice and finally to automated completion of sale.  There is a strong move from capabilities in data mining to the development of capabilities in understanding and interpretation.  There is every indication that this area will evolve steadily.  The Internet may prove to be effective in reducing barriers to trade in services, permit less-developed countries to compete more effectively with services elsewhere, provide a `store' for currently perishable services and lead to great productivity gains in services, although also potentially severe restructuring.

Three dimensional worlds, using tools such as Virtual Reality Markup Language (VRML), are growing in use.  The ability to operate multi-user virtual environments is today limited only by a lack of imagination.  Multi-user Internet games such as "Doom" and "Quake" hint at the vast potential of large groups of people acting in a common environment, similar to the Cyberpunk writings of William Gibson.  DERA, MIT and many other institutions are exploring the development of wholly abstract worlds where people can interact with pure thought constructs, e.g.  financial markets.  DERA's Financial Laboratory is exploring the ability to use VRML to distribute throughout a bank its current financial position in relation to the markets as `blobs' interacting with `jellies'.  Abstract strategic evaluations for defence purposes is an analogous problem to the financial markets.  Much research is directed at reducing the barriers between the "person in the loop" and the machines with which they interact.  Three dimensional worlds are likely to remain an important component in integrating human and computing environments.

The Internet is becoming a truly distributed data environment, each web page interacting with others seamlessly across many servers.  New software such as Java and Inferno combined with network computers, such as the minimum reference model proposed by Oracle, presage the truly distributed computing environment.  Distributed communications and data were at the heart of early Internet design.  Distribution of data and redundancy of pathways increased resiliency, e.g.  to nuclear attack.  Distribution permitted applications such as the World Wide Web where the importance of hierarchies in the software application was absent.  Further distribution of software throughout the Internet, e.g.  "applets" of code, will permit processes to be invoked selectively and partially.  The ability to use only a portion of a large software suite may lead to paying for use of only that portion.  Much new software will be designed to operate across networks of computers.  Developers will attempt to port much old software into these new Internet architectures, e.g.  there already exists an Ada-95 compiler that outputs Java code.  There are problems with porting old software, e.g.  synchronicity of parallel applications and exception control.  Further, Internet design philosophies are inherently different from traditional systems, but the demand for conversion is high.

Conclusion

The defence community remains intensely interested in the Internet and its potential.  There is a requirement to remain strong both offensively and defensively.  Understanding how to interact with civil technology is crucial to keeping up with developments.  DERA provides a world-class capability for accessing advice on many leading Internet issues.

References

Robert X. Cringely, Accidental Empires, Penguin, 1992.

"The Accidental Superhighway", Survey of the Internet, The Economist, 1 July 1995.

"Death of Distance", Survey of Global Telecommunications, The Economist, 30 September 1995.

Donna L Hoffman and Thomas P Novak, "Commercializing the Information Superhighway: Are We in for a Smooth Ride?", Owen Manager 15(2), 2-7, Vanderbilt University, March 1994.

"The InfoWorriers", Wired, 23-24, July 1996.

 

Acknowledgements

This paper draws upon the experiences of the Defence Evaluation and Research Agency (DERA).  DERA is the United Kingdom Ministry of Defence's agency responsible for scientific and technical advice for defence procurement.  DERA has over 14,000 staff and a turnover in excess of UK£1 billion (US$1.5 billion).  DERA has a substantial business, approaching 10% of its turnover, working for commercial and other government bodies outside the Ministry of Defence.  Although the views expressed are mine personally.  Many people in DERA contributed to this paper.  For their help I thank them, but the responsibility for interpreting numerous activities underway throughout a large and dynamic organisation rests with me.  Any misunderstandings are mine.