Table of Contents    Chapter 8    Netguru.net    Glossary

Chapter 9 - Future Trends in Networking

It is always difficult to predict the future. We don't really know what advances will revolutionize our industry within the next few years. However, we can at least identify some broad trends and directions that have been taken. These are likely to be followed in the coming years.

One of the perceived problems of dealing with modern networks is the fact that users sometimes have to be concerned with which network they are dealing with and whether or not they have the right tools to access that network. Currently, if you as a NetWare user decide that you want to communicate with someone who is on a network using the LAN Manager NOS, you would not be able to just hook up networks and trade data. You would have to find a way to translate from one language to the other. You may decide to use a common language to both systems like TCP/IP.

Figure 9-1: The future will hold transparent access to numerous network resources.

If you decide to use a particular printer, some networks may require you to know the location and type of printer as well as what spooler system is handling it and the name of a spooler file or directory (i.e. NetWare queues). You would also have to know in many instances which machine out there among the many you are connected to, contains the application or data you require.

In an ideal networking environment, users should not have to worry about such things. Think about it. The goal of any computer system is to increase productivity. How can you increase someone's productivity if you burden them with overhead like deciding how to get to network resources.

Our industry is concerned about this and has made steps toward achieving a greater degree of transparency for the user. The Consultative Committee on International Telegraphy and Telephony (CCITT) has adapted global messaging and directory services which, when implemented by various networks, will allow users to effortlessly trade mail and access resources more seamlessly. For instance, you may, in the near future, be able to look up the "address" of another person in most any company, and then send a message to them. By the same token, you may be able to send an important document to someone by selecting their printer from a directory and just sending the document there. In addition, you probably will use applications in the future without any knowledge of where the application is located. It might be running on a machine near you, one across the country, or on more than one machine at a time. This transparent ease of access is the goal of modern networking.

The technologies supporting transparent networks are called "distributed computing" and "distributed management".

Figure 9-2: Applications may run more efficiently when their processes are run on several processors simultaneously. This is distributed computing.

Right now programmers generally write applications for specific systems. DOS programmers write DOS programs and UNIX programmers write UNIX programs. Only within the last few years have any honest attempts been made to completely provide real access from one of these OSs to another. Though the barriers may have been political, the future problems are going to be market-driven. The marketplace is demanding universal access.

Distributed computing is the next logical step beyond OS compatibility. The ability of one OS to use the services of another is made possible through Remote Procedural Calls (RPC). RPC provides a universal way of translating a request from one OS to another. Both Novell's NetWare and Sun's NFS use RPC technology. The next goal is to allow the services of various OSs to be used at once by a single application. Not only is this feasible, it is likely. Since RPC allows multiple language translation, we could be running an application that routes some requests through a machine running UNIX while routing other requests to a machine running DOS. The net result is the application runs faster because it does not have to wait for resources of one OS or one machine to be available. Truly the processing becomes distributed.

The final goal of distributed computing is to allow it to take place without the user having to worry about it. All the user should have to worry about is executing Program X. The systems in the background would take the role of finding the application, retrieving it to local memory, and servicing its requests with whatever available resources there are out there. Novell has announced its intention to release a universal set of programming interfaces that will allow clients to be created to run on most all modern network operating systems including the Windows NT product from competitor Microsoft.

The transparency of applications for everyday users is not the only direction for the future. What about managing systems that are so integrated? The management requires a universal way to collect, process and display management data. It also has to include a universal way to manipulate network resources so that data flowing into problem areas can be re-routed. That is the goal of protocols such as Common Management Information Protocol (CMIP) and Simple Network Management Protocol (SNMP). Vendors provide management software and hardware that use these languages. IBM is currently supporting a concept called DME or Distributed Management Environment. The goal of these programs is to extend control over any device in a network from mainframes to PCs.

Most modern management interfaces are graphical in nature, displaying at a glance the status of network resources and traffic. Most work under the premise that there is some management agent residing on a network device, gathering data that it sends to a central management facility. This central facility has built in alarms that alert management personnel about problems or if some element of the network has slipped outside of acceptable limits. Depending on the complexity of the system, it may attempt to correct or circumvent the problem automatically based on certain guidelines provided to it, or the system may require direct operator intervention to correct problems.

We would probably be surprised if we realized how often we interact with some network today. We talk on telephone networks. We do banking on bank networks. We watch TV on broadcast or cable networks. The list of interactions in a day could grow with emerging technology.

Outside of the realm of straight computer data usage, we have the rise of interactive television expected to reach us on a decent scale soon. Several companies have been pioneering this technology which involves the transmission of video and audio data to a user that has an interface allowing him or her to select items by sending an electronic signal back through the cable system to a particular location. The home shopping channels may never be the same.

Data for computers will find new users by becoming less limited by the medium that must carry data. Wireless technology is in full swing now. Infrared and radio vendors are battling it out in the marketplace, and there is no clear winner as yet. Each has its own advantages and disadvantages. There is quite a polarization going on in this area with many vendors carrying one or the other of these systems as opposed to offering both.

Recent FCC rulings have allowed bandwidth to be available for PCDs (Personal Communication Devices). This is going to open the realm of portable handheld computers (such as Apple's Newton) with such powerful features as faxing capabilities and electronic mail access. In a wireless world, we can finish a report while outside of San Francisco, download or fax it to colleagues in Dallas, where it is edited and eventually routed to our publisher in New York. Portability will be key and access will be about everywhere.

There is another factor in the evolution of rapid data transfer. The telecommunications infrastructure must be capable of supporting voice and data transit. ISDN is a positive step in this direction, removing many of the old voice line/data line distinctions.

Continued growth in networks will create some changes in the way computer companies and companies with computers operate. The incredible growth in computer networks has created new niche markets and increased demands in more traditional areas.

The clear trends in networking are the integration of all of a company's computer resources into an "enterprise network" and the connection of larger network entities on a global scale called "global networking".

Enterprise networks have created a whole new business involved in assisting management in implementing their system integration. These consultants are brought in to oversee these projects and assist in managing the resultant hybrid network when the task is complete. Going outside of one's own company to bring in independent help is called "outsourcing". Today, outsourcing makes a lot of sense. If, for instance, you decide to do a little upgrading on your network, it may make a lot more since to bring someone in for a short period of time to accomplish the task, and provide training to your people, than hiring and training full-time personnel. After the project, you would still have a person around that you may not find another use for.

Figure 9-3: The future may hold global access for many organizations.

Global networks are often the result of vendor consortiums or government involvement. The Internet, for instance, sprang from the Department of Defense. Global networks bring opportunities of joint research projects between companies. They also help to standardize processes, applications and protocols for future global networks. As these benefits have become more evident, it is no doubt that global networks will continue to expand in the future. The result may one day be a universal network with access to individual networks all over the globe. When this is accomplished, communication with someone in India will be as easy as communication with someone in Milwaukee or next door.

The success of enterprise or global networks will depend on the structures that allow access to entities within them as well as without. The directory structure of X.500 specifications will aid in this access. Novell, Inc. has gambled quite a bit on the future expansion of local networks into enterprise and global entities. The NetWare 4.X product is a radical departure from the server-based and server-centric environment of the past. It is a corporate recognition of the need to quit thinking of networks as a pathway to that file server over there, as opposed to an extension of my local machine to numerous resources where location is unimportant. Although the idea is not new (Banyan has been marketing such a system for years), Novell as the PC-network market leader has thrown its weight behind a concept of enterprise and global network resource distribution and management. Only time will tell if they have moved at the right time or not.

1. Know the trends pointing toward the future of networking.

 

Transparency

 

Distributed Computing

 

Distributed Management

 

Increased User Access

 

Enterprise and Global Networks

 

2. Know how Novell has invested in the future direction of networking.

 

 

 

 

Table of Contents    Chapter 8    Netguru.net    Glossary

Copyright© 1993 & 1997 by Atrium Technical Inc. All rights reserved. Illegal to copy without written permission of ATI.