Network management (NM) has today, matured way beyond what it used to be a few years back. Earlier, network management was looked more from the device management perspective, wherein the tools provided by the network device vendor were used to primarily configure and administer the network. These tools also had some in-built monitoring mechanism, which was fairly adequate.

However, in the last few years, there are clearly two trends that are emerging in the market. For one, network management is clearly heading towards highly integrated management solution sets, with an emphasis on pre-emption rather than reaction. Organizations are increasingly looking at network management as a sub-set of enterprise management systems. This would enable customers to deploy and manage networks based on policies and also move towards ‘on-demand computing’.

Secondly, organizations are considering remote management, through the Internet or through an external managed service provider, to manage complex network infrastructure. These service providers are measured against the service levels committed to and the service fee payable towards the service providers is directly linked to the uptime maintained by the respective service providers.

TECHNOLOGY TRENDS
l Network Management Functions: Organizations deploying various network and communication technologies are often forced to look at the following key management functions, often referred to as FCAPS, to effectively administer and manage their networks. FCAPS, the ISO model for network management stands for

Fault: Network problems or faults are found and fixed.

Configuration: The network is monitored and controlled. This includes keeping track of hardware and software on the network and any modifications to them.

Accounting: Network resources are distributed and used equitably; end users and departments are charged for their network use.

Performance: Network congestion and bottlenecks are minimized. Current trends and capacity planning for future needs are taken care of.

Security: Only the people who really need access to specific network resources are allowed to use them. This applies equally to outside hackers and internal users.

l Functionality Levels: NM tools have four basic levels of functionality, each having a set of tasks defined to manage the objects in the network.

Managed Objects: Devices, or any network element, like routers, hosts, servers require some form of monitoring and management.

Element Management System (EMS): EMS manages specific portions of the network like async lines, multiplexers or EPABXs.

Manager of Managers Systems (MoM): MoM systems integrate the information associated with several EMSes.

User Interface: The entire log of the network monitored is distributed to the entire MIS of an organization.

l Network Management Protocols: There are many network management protocols available. The two mainstream protocols however are SNMP (the Simple Network Management Protocol) and CMIP (the Common Management Information Protocol).

Generally, SNMP works under TCP/IP and CMIP works under OSI.

The information the SNMP and CMIP can attain from a network is defined as a MIB (management information base). The MIB is structured like a tree. At the top of the tree is the most general information available about a network. Each branch of the tree then gets more detailed into a specific network area, with the leaves of the tree as specific as the MIB can get. For instance, devices may be a parent in the tree, its children being serial devices and parallel devices. The value of these may be 6, 2, 4 accordingly; with the numbers corresponding to the number of devices attached (4 parallel + 2 serial = 6 total devices). Each node in the MIB tree is referred to as a variable. The top of a LAN MIB tree is usually the Internet.

l On Demand or Utility Computing: Organizations have to adopt some form of policy-based management tools that will dynamically manage resources based on demand. These are often referred in the industry as ‘managing on demand’ or ‘on-demand computing’. The whole concept will revolve around understanding current demand by various self-learning techniques and re-prioritizing existing activities in such a way that the business does not suffer due to poor infrastructure, while at the same time taking care of immediate needs.

Utility computing rests on different pillars like business agility metrics, virtualization software for automated resource utilization and self-healing software for network management tools that address technical issues before machines go down, a practice also known as autonomic computing. In addition, utility computing makes it possible to virtualize data center resources such as servers, networking elements, storage and applications and reallocate them according to need.