This talk focuses on application of Massively Parallel Processing, MPP for data warehousing. The target audience is not expected to have a background in MPP or very large database design. The talk is based on testing performed for two large financial companies in New England.

MPP, Massively Parallel Processing has moved out of the research labs and is in active commercial use. This presentation will feature a brief overview of how a shared nothing MPP database works, focusing on function shipping vs data shipping. The key issues for MPP database will then be examined including optimal distribution of data across the MPP nodes, data collocation, scalability, and join strategies. Hardware configurations will be discussed, including SCSI disk vs. RAID. The talk is based on usability testing of Informix XPS on 2,4,8, and 16 nodes on an IBM SP2 and tests on a Sun PDB Cluster.

People will learn:

What is an MPP database?

Multiple computers need to work together on the same database query. We'll examine how they are interconnected with a high speed network and the role the network plays. The software layer needs to be able to distribute the problem to multiple CPU's and manage all available resources in parallel.

How do the individual processors work together?

The two most common distributed database architectures are shared nothing with function shipping and using a distributed lock manager. These two approaches will be explained as well as the benefits and limitations of these approaches. Shared nothing scalability is the primary difference between the two architectures.
Horizontal and vertical parallelism will be motivated by a review of the four components needed: data partitioning, encapsulation of function, dataflow, and control.

What are the challenges to the MPP architecture?

Early MPP database implementations suffered from several problems. Scalability: when additional nodes are added to the MPP environment, what speedup/scaleup is achieved? Collocation: Data is spread across all the nodes in the MPP environment. What penalty is paid when a row of data needed for a join is not local to the node performing that join? I/O restrictions: large database systems historically have input output bottlenecks. Large data caches in the disk subsystem are proposed as a solution. The testing results will be a surprise.

What test results were obtained when testing for these challenges?

Tests on an IBM SP/2 with 2,4,8, and 16 nodes was performed to characterize the above issues. Results for a Sun PDB cluster will also be presented.

The discussion will conclude with recommendations for setting up a data warehouse environment. Can a "production" environment be maintained when database queries to a warehouse are by definition developed on an iterative "ad hoc" basis.

Speaker Biography

Marty Lurie is a Systems Engineer in the Informix Boston office. His focus is applying database technology in the financial sector for DSS and OLTP applications and benchmarks. He has worked with SMP, SMP Cluster, and MPP systems in 2,4,8,16, and 32 node configurations and has tested scalability, collocation, skew and other MPP parameters. Prior to Informix, Martin spent 17 years at IBM in Systems Engineering, development, marketing, and finance. Martin holds a Bachelor's degree in EE from MIT and an MBA from Boston University. He is a Certified AIX Administrator, Certified AIX Instructor, and is Certified in Production and Inventory Management.

May Meeting

Subject

Developing Groupware Applications for the Web using Lotus Domino

Speaker

Rizwan Virk
CEO, Brainstorm Technology Inc.

Date

Thursday, May 15, 1997

Time

Light refreshments and informal discussion from 6:30 to 7:00 PM.
The formal part of the meeting starts at 7:00 PM.

Location

The Newman Auditorium at Bolt Beranek and Newman (BBN), 70 Fawcett Street, Cambridge, MA

Meeting Overview

This talk will focus on using Lotus Domino as a development and deployment environment for building web-enabled groupware applications. Domino is the latest server of Lotus Notes and is an HTTP server and a Notes server that communicates with multiple clients. Lotus has integrated all of the features of Notes and brought them to the Web.

As the focus in the web world shifts from creating company marketing web sites to building more interactive inter/intranet applications, integrated development environments like Domino will become more important.

This talk will give an overview of how Domino applications are built, and will delve into several important areas for application development and deployment, including:

Security with Domino
Distributed Authoring for Web Sites
Creating and Submitting Forms
Initiating Server-based agents
Planning a Groupware Application: Domino User Model
Writing code for Domino applications
Replication and other key features of Notes

Speaker's Biography

Rizwan Virk is CEO and co-founder of Brainstorm Technology, Inc, the leader in groupware products and services. Rizwan has been working with Lotus Notes since 1992, and is a recognized expert in the industry on matters concerning groupware, the internet and enterprise integration. Rizwan is a contributing author to Lotus Notes 4 Unleashed, and was the author of a monthly column in Databased Advisor about Lotus Notes development. He is the editor of a soon to be published book, Domino Survival Guide. Rizwan has given presentations at numerous industry events, including Lotussphere, the Lotus European Technology Conferenece, the WorldWide Association of Lotus Notes Users, the Boston Notes User's Group, and the Boston WebMaster's Guild. Rizwan has consulted companies around the world on the use of groupware technology and enterprise applications. Prior to starting Brainstorm, Rizwan consulted at a variety of companies internationally, including Lotus Development Corporation, Fidelity, and KLM. Rizwan holds a B.S. in Computer Science and Engineering from the Massachusetts Institute of Technology and can be reached at rvirk@braintech.com.

April Meeting

Subject

The Realities of Web Performance and Usage Analysis

Speaker

Matthew Cutler
Founder and Director, Business Development
net.Genesis Corp.

Date

Thursday, April 17, 1997

Time

Light refreshments and informal discussion from 6:30 to 7:00 PM.
The formal part of the meeting starts at 7:00 PM.

Location

The Newman Auditorium at Bolt Beranek and Newman (BBN), 70 Fawcett Street, Cambridge, MA

Meeting Overview

This talk provides a technical overview of Web site traffic analysis and performance monitoring. Material covered includes fundamentals of log file interpretation, primary factors affecting site performance, and practical applications of targetted analysis. Presented by an established industry leader, this session provides clear understanding of the increasingly complex field of site tracking.

Speaker's Biography

Matthew Cutler, net.Genesis Corp.'s director of business development, helped found the company in January of 1994. Today, he leads net.Genesis' strategic product planning and marketplace education efforts. In addition to this role, Mr. Cutler also serves as President of the Webmasters' Guild, the country's first professional association of Webmasters. He frequently presents at major national and international Internet-related conferences and is a contributing columnist to Webmaster Magazine and Internet World Magazine. Prior to forming net.Genesis, Cutler performed basic and applied orthopaedic biomechanics research at Boston's Beth Israel Hospital, and earlier served as an Information Technology Analyst in Aetna Life & Casualty's Bond Department. Cutler graduated Tau Beta Pi from the Massachusetts Institute of Technology with a Bachelor of Science in Mechanical Engineering.

March Meeting

Subject

Distributed Computing in Java

Speaker

Ken Arnold, Javasoft

Date

Thursday, March 20 1997

Time

Light refreshments and informal discussion from 6:30 to 7:00 PM.
The formal part of the meeting starts at 7:00 PM.

Location

The Newman Auditorium at Bolt Beranek and Newman (BBN), 70 Fawcett Street, Cambridge, MA

Meeting Overview

Ken Arnold will discuss distributed computing in Java, specifically covering the Java Remote Message Invocation system (RMI) and object serialization, both part of Java 1.1. He will also discuss JavaSpaces, a distributed protocol development tool, which is currently under development.

Speaker's Biography

Ken Arnold is co-author, with James Gosling, of The Java(tm) Programming Language, part of the official Sun series of books on the Java language, packages, and environment, published by Addison Wesley. He is a leading expert in object-oriented design and implementation, and has written extensively on C and C++ topics for UNIX Review, and is also co-author, with John Peyton, of A C User's Guide to ANSI C.

Ken is a Staff Engineer with Sun Microsystems, previously in Sun Labs, and now in JavaSoft. His is currently part of the team developing the JavaSoft Remote Messaging Interface for communication between Java code running on different machines.

Before coming to Sun, Ken's experience includes being part of the original Hewlett-Packard team designing CORBA, several user interface and UNIX projects at Apollo Computers, and molecular graphics at U.C. San Francisco. In olden days, he was part of the 4BSD team at U.C. Berkeley, where he created the curses library package for terminal-independent screen-oriented programs, and was co-author, with Mike Toy, of the computer game rogue. He received his A.B. in Computer Science from U.C. Berkeley in 1985.

February Meeting

Subject

Gauging the True Work Effort While Eliminating Defects

Speaker

Paul J. Newcum

Date

Thursday, February 20, 1997

Time

Light refreshments and informal discussion from 6:30 to 7:00 PM.
The formal part of the meeting starts at 7:00 PM.

Location

The Newman Auditorium at Bolt Beranek and Newman (BBN), 70 Fawcett Street, Cambridge, MA

Meeting Overview

In general, we gauge the true work by knowing the quantity and work efforts of fully creating useful individual pieces we need for the project. The individual work efforts must include an amount of time to build the item, to eliminate defects, to handle the people communications, to utilize development tools, and to test each item. A local quantity database must record these throughout the project. These individual quantities are fed to a local spreadsheet to calculate the total work effort for everything. These are fed to a local project management package to display meaningful work effort, cost and schedule charts.

Even before calculating those totals, we first begin with just the requirements. We quantify each up front 'plain-speak' business requirement statement. These are simple verb statements that capture the business requirements. These are needed to gauge the Amplification of Functionality Factor, which recognizes that for every individual business statement, literally hundreds and sometimes thousands of technical statements are needed to create each one in the system. Using this factor with the known simplified business requirements statements produces an answer that can accurately predict up front all true total work, the full time frame, the final cost and even the risks of creating and delivering effective computer-functions. As the project progresses, we can iteratively compute this factor and the total amount of project work just from these business requirement statements alone. We can easily see the quantity of work remaining and the time remaining by counting the number of 'plain-speak' business requirement statements and multiplying them with this factor.

Next, with the technical design we quantify the number of distinct user screen (including browsing, navigation, inquiry and update plus help and tutorial), algorithm and database functions in the system. Be sure to account for data clean up and data transport work to the new system. Also account for the help, tutorial and production preparation work. Then, we quantify the number of distinct tests to perform and the work for each distinct type, as we have planned them in our testing approach. Finally, for each distinct automated development tool function we will use on the project, quantify the distinct work to create workable functions and then eliminate defects with this tool.

For each of these, a known creation work effort, a known defect detection and elimination work effort and a known people communication and coordination work effort based upon real world actual experience is needed.

Plug these quantities into a spreadsheet to total up all the creation work, all defect detection and elimination work and all the people communication work per distinct type.

This numeric data then goes into your preferred project management package to build a realistic project schedule in Pert and Gantt chart format.

As the project proceeds, we quantify the number of defects we find and the work effort to eliminate them, plus any additional functions, tests and automated tools appearing on the project. These quantities 'update' our gauge of the true work effort, the costs and the schedule.

Using this data, we can accurately determine how much work there is, the schedule and the complexity of work either based upon the total number of business requirements statements, or based upon the technical design for screens, algorithms and database functions. Thus we have two ways to predict the true work, cost and time frame of the project. Both ways are needed, to help us evaluate the project.

As the work progresses, we want to eliminate defects and errors. Reviews accomplish this. The established requirements and test plan illuminate what the system 'should look, act and perform like'. We review components to find defects, to eliminate them and to guarantee what we are creating matches the requirements.

The developer verbalizes and explains the created item. During this explanation, defects and errors are spontaneously detected and eliminated as they are self-discovered by the developer, and by the team.

Before finishing the review, the developer and the team look again at the requirements to verify that the created components meet them. Once the review is completed, the quantity of defects detected and eliminated are fed into the local quantity databases discussed here.

Therefore, we have a continuous process which accurately gauges the true work, cost and time frame of a project while we are eliminating defects and errors.

The final project should come in on time, for a planned budget, with nearly zero defects. Everyone should be pleased!

Speaker's Biography

Paul J. Newcum has over thirty-plus years of experience assisting people to improve business while achieving quality information systems and better performance. He has been particularly effective in developing solutions in less time and at lower costs than his competitors and peers have achieved under similar circumstances with comparable requirements.

His undergraduate education was in business, and he particularly is interested in helping business achieve tangible results with their information systems. He has direct hands-on skills in up-to-date management and programming arenas including client/server and distributed computing. He teaches others how to achieve significantly more information system effectiveness than are realized by most.

Paul employs a business oriented vocabulary rather than incomprehensible jargon-speak. He communicates issues in language managers appreciate, rather than technical jargon. He frequently helps information systems technicians overcome the inherent communications gap between top management. His goal is to help information systems people speak plainly, while accurately conveying required information.

Paul is noted for an exceptional information systems technique that can reduce the time and effort of projects. What would normally take years of software development can be reduced to much shorter time frames.

Paul began his career by serving in the Department of Defense as a programmer analyst, leading a team to deliver analytical information to the top brass. This work resulted in an invitation from IBM to join Big Blue's software development and support teams. While at IBM, he cultivated practices that resulted in providing effective information systems with exceptional ingenuity and accuracy. As a result of his unique methods, he was able to achieve better results even with fewer people and less resources than were optimally suggested. IBM's clients were extremely pleased that their requirements had been met, the result being that his team received an A+ grade on their information systems. He devised and mastered these more efficient techniques from the very beginning rather than wasting time on glitzy technical features.

During the late 1970's, the computer industry endured immense technical and cultural changes, the conclusion of which was the birth of modern IBM mainframes as well as the beginnings PC's. At this time, Paul formed his own consulting firm to pilot others into these new kingdoms.

Today, he assists people to achieve more in less time, with greater business effect; he assists companies to build quality information systems solutions. He helps professionals grasp the importance of creating modular, reliable information systems. For a long time, he has been teaching information systems people to incorporate 'best practices' into their work.

Paul has personally delivered hundreds of successful information systems. He has also designed, written, tested, and satisfactorily delivered well over ten-plus million lines of programming to assorted arenas, including: nearly every industry, modern relational database business systems, and software tools for professional developers. He defines the causes of and the corrections for prevalent software defects and errors.

Paul is published in a variety of national professional publications as well as corporate newsletters. Several organizations confer with him to organize and present their information on the INTERNET and the WORLD WIDE WEB. He holds memberships in the ACM, ADAPT, ASM, ASQC, IMC, and SIM. He stays abreast of the most effective solutions for information systems.

Paul possesses an undergraduate degree in business and programming from the University of the South, and he subsequently obtained a MS as well as a PhD in computer science at the Pacific Western University, a distant learning center for graduate studies and research. Additionally he gained learning from a number of technical centers, among them: IBM, Microsoft, Oracle and others.

January Meeting

Subject

The Paradigm Shift from Algorithms to Objects and Interaction

Speaker

Dr. Peter Wegner
Brown University

Date

Thursday, January 16, 1997

Time

Light refreshments and informal discussion from 6:30 to 7:00 PM.
The formal part of the meeting starts at 7:00 PM.

Location

The Newman Auditorium at Bolt Beranek and Newman (BBN), 70 Fawcett Street, Cambridge, MA

Meeting Overview

The paradigm shift from algorithms to objects and interaction captures the technology shift from mainframes to workstations and networks, from number-crunching to embedded systems and graphical user interfaces, and from procedure-oriented to object-based and distributed programming. Interaction is shown to be more powerful than rule-based algorithms for computer problem solving, overturning the prevalent view that all computing is expressible as algorithms. The radical notion that interactive systems are more powerful problem-solving engines than algorithms is the basis for a new paradigm for computing technology built around the unifying concept of interaction.

Speaker's Biography

Peter Wegner, a professor of computer science at Brown University, was educated at London and Cambridge Universities, and taught at Cornell, Penn State, and the London School of Economics before coming to Brown. His research interests include programming languages and software engineering. His books include the first book on Ada and edited books on research directions in software engineering and object-oriented programming. His research has included language design, concurrency, and type-theory issues in object-oriented programming. He is currently exploring component-based software technology as an integrating multiparadigm framework for object-oriented, distributed, and database technology.

December Meeting

Subject

Why Surf Alone? Intelligent Agents to Help You Explore the Web

Speaker

Henry Lieberman
MIT Media Lab

Date

Thursday, December 12, 1996 (Note: Second Thursday)

Time

Light refreshments and informal discussion from 6:30 to 7:00 PM.
The formal part of the meeting starts at 7:00 PM.

Location

The Newman Auditorium at Bolt Beranek and Newman (BBN), 70 Fawcett Street, Cambridge, MA

Meeting Overview

The next major development in the evolution of the Internet will be introduction of "intelligent agents", programs that assist you in using interactive interfaces for information search and other tasks. These agents will be running independently of the user's activities, and will be capable of learning from interacting with the user and providing help or suggestions as to what to do next. I'll illustrate this by showing Letizia, an agent that assists Web browsing by providing a "channel surfing" window that continuously displays recommendations. It learns the user's preferences by recording the user's choices in the browser and searches the "neighborhood" of the current page for other pages of interest. It treats Web browsing as a cooperative search activity between the human user and the computer agent, providing a middle ground between narrowly targeted retrieval such as provided by search engines, and completely unconstrained manual browsing.

Speaker's Biography

Henry Lieberman has been a Research Scientist at the MIT Media Laboratory since 1987. His interests are in the intersection of computer graphics, human interface, and artificial intelligence. His current projects involve media interfaces that learn from examples presented by the user. He is a member of the Learning and Common Sense Group, which works on the idea of interface agents, intelligent assistants for interactive media applications. He has also worked with the Visible Language Workshop group, which is concerned with visual design issues. Prior to that, he was a researcher at the MIT Artificial Intelligence Laboratory since 1972, where he worked on parallel object-oriented programming, knowledge representation, programming environments, machine learning, and computer systems for education. He holds a doctoral-equivalent degree from the University of Paris and was a Visiting Professor there in 1989-90. He has published over 40 papers on a wide variety of research topics.

November Meeting

Subject

Distance Learning and Collaboration Using Multiple Technologies

Speaker

Dr. Lisa Neal
EDS Center for Advanced Research

Date

Thursday, November 21, 1996
(Note: Misprint in Real Times newsletter: Meeting is Thursday, not Monday)

Time

Light refreshments and informal discussion from 6:30 to 7:00 PM.
The formal part of the meeting starts at 7:00 PM.

Location

The Newman Auditorium at Bolt Beranek and Newman (BBN), 70 Fawcett Street, Cambridge, MA

Meeting Overview

Organizations are searching for cost-effective ways to reduce the time and expense of travel. This is especially apparent with education, since it is traditionally offered in a face-to-face classroom. We experimented with the use of multiple interaction technologies to provide an effective alternative to face-to-face instruction. This talk describes how various communications technologies, including videoconferencing, audioconferencing, Internet Relay Chat (IRC), email, and a corporate intranet, were employed to teach courses to geographically-dispersed participants. The same technologies were also used for collaborative projects.

The talk will cover the selection of distance learning and the delivery technologies, deployment strategies and issues, and participant feedback. We found that the use of multiple technologies provided richer communication than any one technology alone. Not only was the course content covered, but the subtler aspects of classroom interaction were supported, such as forming relationships with other participants. However, the effective use of each technology was far more complex than we had imagined at the outset. We also found that we were able to reach more people than we would have been able to without distance learning. This had ramifications for other classes and other instructors who were interested in increasing the number of students they reached, without decreasing the impact of the material.

Speaker's Biography:

Lisa Neal is a senior research engineer at EDS, where she develops and delivers courses on Human-Computer Interaction (HCI) and Collaborative Environments and also consults in these areas. She is researching the effectiveness of collaborative tools for distance learning. Lisa previously worked with the Capture Lab, EDS computer-supported meeting room, and researched the organizational impact and productivity gains resulting from the use of computer-supported meeting rooms. Prior to joining EDS, she completed a Ph.D. in Computer Science at Harvard University, where she also was a postdoctoral fellow.

Lisa is workshops chair for CSCW'96. She has presented numerous tutorials and seminars, most recently a seminar on Emerging Technologies in HCI at the Technical University of Darmstadt, Germany in October 1996. Lisa is co-editor of "Structure-Based Editors and Environments," Academic Press, 1996.

October, 1996 Meeting

Subject

Building Virtual Teams with Computer Mediated Communication Systems

Speaker

Dr. Merrill E. Warkentin
Associate Professor of MIS, Northeastern University

Date

Thursday, October 17, 1996

Meeting Overview

Organizations are forming virtual teams of geographically distributed knowledge workers to complete workplace tasks. Various computer-mediated communications systems (CMCS) have been developed to facilitate effective collaboration between team members at remote sites. These various systems will be briefly presented and categorized.

At Northeastern University, we have developed a CMCS called "MeetingWeb" (tm), an asynchronous computer conferencing system which provides eponymous textual communication capabilities to its users. MeetingWeb is a moderated web-based bulletin board which allows participants (who are issued passwords) to post topics and comments anytime day or night, thereby facilitating team collaboration. The moderator can "seed" the conference with interesting or provocative topics. Participants may also begin their own topics -- they can post documents online to be read by others. Or a user may simply read the comments of others without contributing. This creates a 24x7 virtual discussion in which anyone can participate even if they are normally shy.

We initiated a project to create virtual teams composed of MBA students at Northeastern University and "managers in training" at other universities around the world. The teams engaged in several research projects for which they were tasked with using the web search engines to find information pertaining to assigned topics related to Information Technology. The students at both universities were encouraged to organize their teams and to assign responsibilities via the virtual conference or via email. The final reports were posted on the conference online rather than submitted on paper! And the reports contained actual hot hyperlinks to the web sites they uncovered in their research. Their work was also graded online with temporary easy detours to the websites they placed (linked) into their reports.

An instrument (consisting of 26 items) was designed to assess students' experience with the system, including a thirteen-item satisfaction measure specific to Computer-Mediated Communication Systems (CMCS). Results of this research will be presented and discussed. It was determined that organizational factors such as "social presence," balanced composition, and training have a greater influence on outcome than technological factors. We will also explore the communication process of remote groups who collaborate using CMCS. Finally, guidelines for creating and managing virtual teams will be presented.

Speaker's Biography:

Merrill E. Warkentin is Associate Professor of Management Information Systems in the College of Business Administration at Northeastern University in Boston, MA. He has authored over 80 articles, chapters, and books, including most recently, Emerging Information Technologies. Professor Warkentin's research, primarily involving information technology management, artificial intelligence, computer security, digital multimedia, and the World Wide Web, has appeared in such journals as MIS Quarterly, Decision Sciences, Expert Systems, AI and Medicine, The Journal of Knowledge Engineering & Technology, PC AI, Journal of Computer Information Systems, Applied Computing Review, and The Journal of Intelligent Technologies. Dr. Warkentin has served as a consultant to numerous companies and government agencies, including the United Nations, the U.S. Navy, the Internal Revenue Service, the Egyptian Ministry of Agriculture, the Kenyan Ministry of Finance, the Nebraska State Patrol, MCI, and SRI/Gallup. He has also lectured at the Army Logistics Management College and has been a featured speaker at dozens of association meetings and industry groups. He holds BA, MA, and Ph.D. degrees from the University of Nebraska-Lincoln. You can reach Professor Warkentin at (617) 373-2044 mwarkentin@neu.edu. His URL is http://www.cba.neu.edu/~mwarkentin

September, 1996 Meeting

Subject

Emerging Wireless Data Communications Systems

Speaker

Craig J. Mathias
Principal, Farpoint Group

Date

Thursday, September 26, 1996

Meeting Overview

Over the past five years, the emphasis in wireless communications has seen a rapid shift from fixed to mobile systems. Despite the increasing importance of fiber as a backbone for long-haul networks, fixed wireless remains a viable competitor, achieving throughput that can match fiber at up to OC-3 speeds (155 Mbps). But mobile systems dominate current interest, driven by a number of factors. These include an increasing mobile workforce demanding "anytime, anywhere" connectivity, the difficulty of providing a wired infrastructure in emerging economies, and, most interestingly, a significant interest in wireless data communications, driven by the rapid growth of the installed base of mobile computers and emerging combined voice/data devices.

Many engineers are familiar with the challenges of designing, building, and configuring and installing wireless systems. Mobility adds yet another dimension, in that traffic engineering takes on a certain "statistical" quality that often results (as many cellular-telephone users are aware) in noisy connections and even dropped calls. While all-digital wireless networks can help in this area, they are not a universal solution. And data communications is even less forgiving, since throughput is always of paramount concern and the airwaves represent a potentially hostile medium. The combination of limited bandwidth, systems originally designed for voice communications now being retrofitted to handle data, and the vagaries of radio propagation all combine to create a most challenging environment, but one which must be addressed to meet the needs of very rapidly-emerging market demand.

Despite all this, great progress has been made in wireless data. Throughput on circuit cellular connections, using appropriate modems, can exceed in some cases 20 Kbps, and emerging digital systems hold the promise of even higher throughput - systems offering .5 Mbps into a handset-like device have been described. Moreover, wireless LAN systems, offering in-building mobility, have broken out of the 1-2 Mbps range and some systems now offer 3 to even 10 Mbps of throughput. With recently-proposed changes to the regulatory environment, wireless LANs with even 100 Mbps may not be far off.

Wireless data is moving rapidly from an adjunct to voice-oriented systems to networks designed exclusively for data communications, on a scale ranging from in-building to world-wide. Despite the traditional problems inherent in wireless communications, data is rapidly finding a home across a broad scale of applications and markets.

This session will provide an overview of the key issues of wireless communications, with a focus on data. A basic understanding of the issues of RF communications will be assumed for this talk, although a brief review will be provided. The emphasis will be on emerging systems, including wireless LANs, Cellular Digital Packet Data (CDPD), and Personal Communications Systems (PCS). We will take a brief look at each of these, and discuss their key characteristics and capabilities. Some speculation as to the future direction of broadband wireless will also be provided.

Speaker's Biography:

Craig J. Mathias is a Principal with Farpoint Group, an advisory and systems-integration firm based in Ashland, MA, specializing in emerging communications technologies. The company provides technology analysis, strategy development, program management, and product specification, design, and marketing services for manufacturers, and system planning, deployment, business-process re-engineering services for end-users. Craig is a recognized expert on wireless communications and mobile computing, and is a frequent speaker at conferences and contributor to major journals. He is a member of the IEEE, and holds an Sc.B. in Applied Mathematics/Computer Science from Brown University. He is also Chair of the Advisory Board of Wireless LAN Research Labs, at Worcester Polytechnic Institute (WPI).