With the quickened pace of technological advances in recent years, from Telecommunications Technologies Information Technologies, Computing Technologies, to VLSI Technologies, the Telecommunications Networks of the year 2000 will be far more advanced and complex than today networks. That realization poses a challenge for many companies to look for ways to evolve their networks while still being able to manage it. To the Telecommunications carriers and service providers, this is both a challenge and an opportunity to provide new services. 

By year 2000, 3 years from now, most of the technologies that engineers are currently playing with in the labs will be widely deployed in the network. From very high speed access technologies such as ADSL, and Cable Modems, to Multimedia/Multi-Services network based on ATM technologies, to ultra fast servers that can provide instantaneous access to information, all these will make some of current network management methodologies obsolete. Wireless network access will make traffic collection and traffic pattern recognition impossible. As wireless access becomes more widely used traffic engineering and traffic network management will be more difficult. With Multimedia, the definition of network performance will change, methodologies to measure and manage networks for peak performance will need to be revised. 

In this paper we will look at some of the advances in Technologies and how it will shape the Telecommunications networks of the year 2000. With that view, we will present options on how to manage such networks together with the issues involved. 

Access networks are defined as the portion of the network that connects the end-user devices, such as telephone sets, ISDN Terminal Adapters, Modems, etc., to the backbone network nodes. The perception was that this is where the bottle neck is. There are numerous efforts by many vendors to widen this bottle neck. First the modems are getting faster. Analog modems are expected to be replaced by digital technologies toward the year 2000. ISDN would be much more common place based on the effort by the telephone companies. Cable modems will be in the house of a large number of subscribers. To keep their presence, the phone companies will be deploying widely, and at a competitive price, their ADSL technology in the same neighborhood, if not the same household as the Cable modems.  

Wireless, Cellular, Personal Communications System (PCS) will all be present competing for access bandwidth into the network. Today, in the offices, many layers of networks, and network equipment are interconnected in forms of Local Area Networks (LANs). This architecture is moving to the homes. Many households are starting to have several computing devices in their house. Some "smart" appliances have appeared and soon will be networked with the other "computers" in the house. All these devices are being networked and will continue to be networked. A few home networking solutions have appeared on the market but have not been widely used because of wiring and connecting problems. Wireless LAN is the answer and it will be affordable in a few years.

As the access networks become faster and more widely available, the demand on the backbone network becomes heavier. With traffic come from multiple applications, such as telephony, Internet, Entertainment video (movies on demand, interactive games, etc.), Video Conferencing, Multimedia applications, etc., the backbone network will be equipped with multiple platforms connected by very high speed links. The platforms will likely be a mixture of old and new technologies from many different vendors. For example in the telephone network, many of the older telephone switches will still be in service by year 2000. Sitting next to together with those in the telephone companies’ central offices will be video switches, Asynchronous Transfer Mode (ATM) switches, Data switches, ADSL concentrators, Modem racks, Hubs, etc.

In 1984, at the breakup of AT&T, many large companies saw it as an opportunity to build their own network. After a brief period of feeling in control of their own networks, most of these companies start to experience the pain of ownership. Many of them decided to outsource the equipment, the management, and the control of their networks to a third party, most likely, to their telephone companies. This trend will continue because of cost containment and manageability of network resources. 

By year 2000, a large number of companies with private networks will get out of the networking business and concentrate on their core expertise, be it retails, insurance, airline, or healthcare. Third party outsourcing companies will maintain, control, and manage networks for these companies. The network services provided to the customer companies will be in the form of Virtual Private Networks (VPN). These VPNs will have the look and feel of a private network but are actually part of an integrated network with multiple platforms made by multiple vendors. Outsourcing vendors providing this type of services are telephone companies, cable companies, large ISPs, and other networking companies.

The outsourcing companies have been proposing the Hybrid Networking solution to many large customers. In this scenario, the customer companies maintain, control, and manage a portion of their networks and use the public network to complement it as needed. For example, a customer company may decide to manage their own LANs located in all their offices but use an ISP or a telephone company network to connect the offices together. 

There are many variances to this scenario but it seems as it is the best of both worlds. For the customer companies, they do not have to totally give up their staffs or control of their network resources. For the outsourcing companies, they get "a foot in the door" and potentially more business if they perform well.

Managing this type of network is more complicated because of unclear division of responsibilities and possibly incompatible network management systems. Since at least two or more network management systems are involved in this network (e.g., one for the customer part, and one for the outsourcing company part), it is possible that they do not communicate well, their databases are not synchronized, and their management decisions are conflicting.

The network traffic mix will be very different in the year 2000. More video traffic will be traversing the network. The video traffic will come from various diverse applications with different characteristics. The Entertainment video such as movies on demand will be continuous, delay sensitive, high bandwidth, and point to point or point to multipoint. The Push technology will provide video data in form of files send point to point or broadcast. It is delay insensitive, and low bandwidth. The video telephony and conference will be more point to point with medium bandwidth, and somewhat delay sensitive.  

Other applications, such as Multimedia collaboration and Email, Internet telephone, Internet audio, etc., contribute their traffic with their own characteristics. These data will greatly change the traffic mix in the networks that is traditionally mostly voice or mostly data. The method of managing networks, engineering networks, and forecasting demands will have to be revisited. We used to know how much traffic comes from each city, each town, or each neighborhood based on the number of telephone lines installed there. With the proliferation of cellular and PCS technologies, and then the number portability requirements from the customers, the traffic assumption will have to be changed. The customer may be registered for his cellular phone in Georgia but originate his calls in California. The customer may be ordering her phone number in New York but she can take her phone number with her when she moves to Florida. Traffic volume will be unpredictable, thus making the engineering of network equipment much more challenging.

As we can see, the network to be managed is going to be very complex. The network will be much faster, and the traffic unpredictable. The equipment will be from multiple vendors using different technologies to support a multitude of applications. This is a real challenge for network managers in the year 2000.

There are 5 areas of Network Management: Fault Management, Security Management, Configuration Management, Performance Management, and Accounting Management. Another network management area that has been receiving much interest recently is Customer Network Management (CNM). This is when the customers are given the opportunity and tools to manage their part of the networks. We will discuss these Network Management areas in more details and present options and issues involved with managing networks in the year 2000. 

For large and complex networks, network management exceeds human capability to monitor and react to any event in the network in real time. Network Management Systems (NMS) running on high speed computers will have to be built to keep track of the status of the state changes in the numerous devices, and facilities, and protocols, and users, and applications in the network. The NMS have to be smarter, more scaleable, easier to use, and more adaptable to the changes in networks and technologies. For the following sections, we will discuss the capabilities of the NMS for the year 2000.

Fault management is the process of locating and correcting network problems (faults). Fault management consists of identifying the occurrence of a fault, isolating the cause of the fault, and correcting and tracking the fault.  

When a fault happens in the network, generally a change of state happens in some equipment. One or more alarms are generated. Ideally, we have to capture the alarms and filter them to determine the severity, the type, the devices or services affected, and even the number of occurrences.

Often a network error results in multiple alarms. A simple example is when a link between two devices is cut, at least 2 alarms are generated, one from each end device. Potentially many more may be generated, from excessive packet lost, to incomplete calls, to sessions disconnected, to application time-out, etc. The amount of data generated from one of these network faults can be overwhelming for a person to look at and will take most people a considerable amount of time to sort out what the real problem is. One of the role of an NMS is to correlate all these alarms to determine their cause.  

Sometime an alarm is not a cause for concern. But if the same alarm happens continuously or consistently over a period of time, then it may indicate a more serious problem. The NMS should be able to count the number of occurrences of an event either over an absolute period of time, or over a sliding interval of time. The occurrences can be of a specific event, or event type, or a combination of several events or event types. 

Some alarms are more severe than others. The NMS should be able to recognize the severity of the alarms and react accordingly. Some alarms can be ignored, with only an increase in the statistics counter. Other service affected alarms have to be dealt with immediately, service personnel have to be paged or dispatched, and management has to be notified.  

Alarms will be displayed on an NMS console in one of several forms. They can be displayed on a colored map of the network where each colored location can determine the severity and the location of the alarm. They can also be displayed on a list sorted by severity, location, type, or customer. They can also be propagated to multiple monitors under different formats for different people depending on their functions.

Security management involves protecting sensitive information found on devices attached to a network by controlling access to that information. Sensitive information is any data an organization wants to secure. Security management protects sensitive information by limiting access to hosts and network devices, and by notifying responsible personnel of attempted or successful breaches of security. Protection via security management is achieved through specific configurations of network hosts and devices to control access points within a network. Access points may include software devices, hardware components, and network media.

An NMS should provide a mean to control the security of the network. At any time there can be two types of users accessing the NMS, the customers and the operators. The customers access the NMS to retrieve information about their network usage, their profile, and their connectivity. The customers login profiles should provide them with limited access to their applications and their partitions of the network. Since several competing customers can use the same network provider. It is very important for this provider to keep their data separate and only viewable by them. 

The operators should be organized into several hierarchical levels depending on their authorities and responsibilities. The operator levels can be setup in their login profiles. The authority profiles will determine which operator can access to what system, data or applications. Only the NMS administrator can change these levels of authority. 

Any successful or attempted security violation based on system or connectivity should be reported and analyzed. Security policy should be reviewed and updated often to avoid gaps and weaknesses. 

Configuration management is the process of obtaining data from the network and using it to manage and set up network devices. It consists of 1) gathering information about the current network configuration, 2) using the data to modify the configuration of the network devices, and 3) storing the data, maintaining an up to date inventory, and producing reports based on the data. 

In large and complex networks, it is a voluminous job to keep track of the inventory of the equipment, devices, facilities, together with their locations, network addresses, software versions, maintenance schedules, etc. If a device, such as a switch, fails, a replacement switch is brought in, it is of paramount importance to bring the new switch up to the same software version of the old switch, using the same switching table, the same address, and the same connectivity. This has to be done for the network to function in the same manner as before.  

When network upgrade is necessary, especially software version upgrade, then the NMS with this configuration management feature will become invaluable. It can save countless hours of frustrated troubleshooting to bring the network up and functioning again.

Performance management insures that a network remains accessible and uncongested for maximum efficiency. Performance management should monitor network devices and associated links to determine utilization and error rates. It should also insure that capacities of devices and links are not overtaxed to the extent of adversely impacting performance. 

Performance management should consist of: 1) collecting data on utilization of network devices and links, 2) Analyzing relevant data to discern high utilization trends, 3) setting utilization thresholds, and 4) using simulation to determine how the network can be altered to maximize performance. 

The NMS must provide a process or processes that can collect many types of network performance data. Some of these types of data is: 

• Response time: time it takes for a datum to enter the network, be processed, and leave the network.
• Rejection rate: percentage of time the network cannot transfer information because lack of network resource and performance.
• Availability: percentage of time the network is availability for use. Often measured as Mean Time Between Failure (MTBF).

The NMS should be able to collect performance data on demand or on schedule. Collection time should be a function of devices, links, and absolute or relative times. Information may be collected for specific devices and links or device and link types.

The NMS should be able to provide performance statistic reports to help the network engineer in analyzing the network. These reports will also provide a long term view of the overall health and efficiency of the network. There are 3 types of performance reports: 

1. The real time operations reports: Provide the snap shoots of some parts of the network at a specific time. Example: Switch usage report for Christmas day.
2. The summary or percentage reports: Provide the general view of the performance of some parts of the network during a certain time period. Example: Average switch usage report.
3. The trending reports: Provide the view of the performance of some parts of the network over time. Example: Monthly Average switch usage report. 

The NMS should also have a process to set performance thresholds for various parts of the network. Thresholds are boundaries within which affected devices will function normally. Violation is when one of these boundaries is exceeded. Violation will automatically trigger procedures to collect data, generate reports, or create an alarm. 

Accounting management is the process of measuring network usage to establish metrics, check quotas, determine costs, and bill users. Accounting management includes the gathering of data about utilization of network resources, the setting usage quotas using metrics, and the preparing of data for billing users for their use of the network.

The NMS should have processes to measure the usage of network resources for each customer. The measurement should be on demand and/or scheduled. Since different applications may incur different costs, the NMS should also be able to measure the network usage based on applications. Example: A user may only subscribe to basic Internet access with applications such as Email and Web access. At ayy time this user may want to communicate with the NMS of the service provider to request in real time a higher priced bandwidth for Video conferencing. The NMS should detect the request and begin measuring the Video traffic when it starts so that appropriate charges can be applied.

The NMS should also allow the administration of usage quotas for each customer. The quotas define the amount of network resources a customer is allowed to use under its contracts. The use of resources beyond the quotas will trigger separate measurements and traffic may be blocked or charged at a different price. Network resources should include the NMS, and other customer support functions.

Billing is a complex process that involves taxes, special pricing, promotions, and various discounting schemes that can change frequently. Billing programs should be left to specialized billing companies to do for those reasons and because they differ from company to company, and from one industry to the next. The NMS should be flexible enough to interface with any or all billing programs, especially non-standard legacy ones. The NMS should provide its billing data in the right format to the appropriate billing programs.

Customer Network Management (CNM) allows the customers access to the NMS with applications to manage their own real or virtual networks. The simplest form of CNM is an application that allows the customers to see information for their parts of the network.

Even in this simplest form, many issues can be raised in a large, complex multi-user network. Issues such as how to ensure that the customer can only see their own data? How to keep the data fresh and up to date? What interfaces to provide to the customer? Etc. Some trends on these issues have emerged and will likely to continue into the year 2000 and beyond. Trends such as a Web or Java user interface, and a standard application to application interface using SNMP.

The NMS should provide better CNM functions for the customers to manage their own parts of the network. For example, without the network operator involvement, it should allow the customers to:

1. change their own network routes as needed,
2. schedule changes or updates to their network services, or
3. request specialized reports on their network usage.

To do this the NMS has to maintain a Customer Network Data Base that contains all customer information pertinent to the network. This database may contain information about ordering, provisioning, trouble handling, operation, maintenance contracts, billing options, as well as information about carriers and equipment vendors. Other informational data that should be included in this database are: customer contacts, vendor services, responsible organizations, types of services used or on order, etc. More specifically, the database must contain the exact equipment that the customer has, where they are located, who made them, what software version do they have in it, what network addresses do they have, etc.

The NMS will maintain all these data for each customer. It must have the ability to add, delete, or modify records in this database. Depending on the services, some of these data can be changed by the customers themselves; some can only be changed by the network operators depending on their security levels.

The CNM should also facilitate the communications between the human operators and the human customers. Communications by faxes, Email, and Web sites are all possible and should be maintained and used.

As technologies' advances are measured in months and days as compared to years as in the past, the communications network is getting more complex and is exceeding the capability of human to manage it in any reasonable manner. The challenge we face in the year 2000 and beyond is to build flexible and feature rich Network Management System that can help us to control and administer the networks that we all depend on so much for our day to day life. The network management techniques and concepts will not change as much by year 2000, but the integration of all these techniques and concepts into an integrated NMS will be the theme of the future. For as the networks get larger and more complex, only cohesive integrated NMS can help an organization masters its applications and its services in a cost effective and efficient manner.