C. M. Yuhas

Expert systems in network management-the second revolution

Expert systems are discussed as the next major evolutionary step in network management. Specific examples are presented that demonstrate the benefits of increased programmer productivity, automating network operations, and identifying and diagnosing network problems. Internal structures are discussed. An expert system, functioning in a live environment, that can learn from experience and modify its behavior is presented. Current technological problems in implementation are reviewed and proposed solutions are given. >

Testing network management software

A fourth wave of computer-based network management systems is here. Experienced network managers worry about the reliability of software. This discussion provides ways of testing software to improve software and system reliability, and it proposes a “reliability footprint” in-process metric to monitor the effectiveness of a test program. Systematic ways of moving from debugging to assuring reliable network management systems are emphasized.

How technology shapes network management

Five areas of technological advancement that affect network management are examined. They are: expert systems and machine learning/induction: fault tolerance; object-oriented databases; facsimile; and ease of use (user friendliness). The following points are emphasized. Machine induction will give network managers the opportunity to understand reasons behind actions instead of just seeing the results. Fault tolerance means no scheduled downtime, freedom from obsessive machine management, enhanced opportunity to manipulate processes, and ease of expansion. When object-oriented data management systems are used in the development of network management systems, development efforts are significantly reduced because the same model is used for both programming and database accessing, and the model is a much better representation of the reality of the network than current data models. A major breakthrough will be the ability to access the same data items from two or more database management schemes simultaneously. Facsimile will experience rapid growth and will be integrated with other services. Ease of use is the central element for the intelligent use of future systems.<<ETX>>

The Vision for Networks and Their Management

The future belongs to those who harness the computer to communication networks. The bringing together of byte people with bit people is essential if this future is to happen. Byte people worry about getting data from networks into computers so that applications software can digest it while bit people worry about getting data from one point to another reliably and quickly. Too often byte and bit people fight about whose work is more important. Both are essential as we move to a telepresence society.

Technology Impacting Network Management

How has technology changed the lives of our network managers? First, let’s consider how technology has changed the network. Just 25 years ago, the network was metallic with analog switches and copper transmission. Today there is sand everywhere. Silicone from this sand makes up the integrated solid-state circuits used in digital switches and in the fiber optic cable rapidly replacing copper. We have been through digital passion and fiber fever and now we are headed for fault tolerant mania. Software is the technology that can make aloof hardware friendly and bring information close. It’s a shame that we often miss the mark.

Software Investment Strategy

ABSTRACTFour factors correlate to productivity in the development of real-time, safety-critical systems: programmer skill, the difficulty of the problem being solved, the customer relationship, and the use of the best tools and technology. These factors produce economies in the proportion of 10:5:1:50 on the cost of a project, which suggests an investment strategy. Recommendations are made for managing software engineers to optimize these factors. Large-scale reuse is the technology of the future. The major lesson is that reuse is most economical when modules are not modified.

Thresholds: “I'd Rather Do It Myself, Bill”

Sometimes we sit, chin-in-hand, and wonder if we are still relevant. Have not all the great truths been expounded? Are not the basic tenets of software development long ago learned? Do we need to say any of this yet again? You’d better believe we do. Linux is here and it’s eyeing Bill Gates’ lunch. It needs just a few things to be a real contender. The movement of servers to this new operations system (OS) and away from Windows NT vindicates the underlying concept of its forerunner, UNIX, which was specific tuning, accessible architecture, interaction with end users, and robustness. UNIX is designed to be infinitely malleable in users’ environments even though, after too many additions of options, this created the drawback of making it difficult to administer. Linux has fewer options, is more robust, and its software is free (though its installation is not, as you will see). The cost difference for a mailbox between Linux, approximately

Life at the Edges

30 per year, and Windows NT,

Just How Integrated Are We

100 per year, is impressive. Although Linux is meant to be visible only to the system administrator, its philosophical origins in UNIX sometimes present a complex array of options and choices to the user. The issue is to make Linux transparent to the user, but not restrict the system administrator’s ability to find and fix problems. Linux is sufficiently robust that the problems most often result from configuration issues, rather than inherent bugs, and are therefore easy to fix, if one is well versed in folklore and none too fussy about consistent human factors principles. For example, the address structure for TCP requires port numbers for particular applications to route messages properly. Is “port” a hardware location? No, indeed! It’s an input queue! This is guaranteed to make end users furiously frustrated but system administrators cozy in the feeling of being part of the exclusive cognoscenti. This poem, which we adapted from an anonymous writer on the web, captures the feeling.

ADSL Is a Flawed Gem

We have just returned from a trip to Wisconsin where we had the opportunity to view several of Frank Lloyd Wright’ s buildings. We didn’ t have any original thoughts about them. Many others have observed that FLW, as he is fondly known, had lots of architectural ideas which revolutionized design perspectives, but that the dear fellow came up a tad short on engineering know-how and technological expertise. As a result, the roofs leak badly, the cantilevered bridges are crumbling away, and the mosquitoes banquet on visitors because FLW loved re ̄ ecting water and despised screens. What did strike us was that the current state of network management suffers from the exact inversion of FLW’ s problem. Technologies abound, but we lack an architectural vision. And just as FLW’ s problems occurred at the edges where building met sky, water, and the gravitational pull of the earth, so too our network problems are occurring at the boundaries of technologies.