Networks are like cars. All you need is a smooth drive with minimum traffic
and no breakdowns. But managing a network may not be as simple as driving a car.
In the multi-vendor multi-technology scenario that exists today, a network
manager has to cope with a number of challenges to keep the network up and
running. But if managed effectively, even a modest network can pay large
dividends by saving time, improving communication between employees, customers,
and suppliers, increasing productivity, and opening new paths to worldwide
information resources.
Network management
Network management as it is seen today, goes way beyond network monitoring
or trouble shooting. The objective is to create a centralized, open, scalable
system with the capability to track, monitor and control resources across an
entire network, with minimum downtime. Improving the predictability and
availability of the network would involve a host of activities:
-
Fault
management: To quickly identify potential network problems. -
Configuration
management: Changing network and user configurations to optimize network
performance and productivity. -
Performance
management: This is done for tracking important network events, projecting
future upgrade requirements and troubleshooting. -
Accounting
management: To track and bill network users for their services and software. -
Security
management: To protect the network from unauthorized access to critical business
data. -
Service
management: A key success factor for service providers.
Tools and topology
A network topology or a single view of the entire network and its
components, to determine equipment status and facilitate network configuration,
maintenance and monitoring, gives the capability to list various objects in the
system including list nodes, wireless objects, paths, endpoints and partitions.
For reliable network services, one needs to have an efficient management tool,
capable of providing full control to the network administrator. This would help
in quickly identifying the malfunctioning equipment in a large network scenario.
TIPS FOR CIOs | ||||||||||||||||||||||||||||
|
||||||||||||||||||||||||||||
|
||||||||||||||||||||||||||||
|
There are various kinds of tools available for network management. Network
management platforms such as HP OpenView, Computer Associates Unicenter, and SUN
Solstice. Each network device is represented by a graphical element on the
management platform’s console. Network devices can be configured to send
notifications called SNMP (Simple Network Management Protocol) traps to network
management platforms.
The simple SNMP-based products do a great job of monitoring at a generic
level. These tools actually go far beyond generic monitoring to provide the
ability to keep the OS on the device up-to-date (all relevant patches applied).
They are useful for monitoring vendor-specific information. The have the ability
to detect changes to the memory on the device for example, changes to
configurations, etc. Also important is their ability to audit changes occurring
in the network along with a trail to the person making the changes. One needs to
be careful on the availability of continued support and enhancements on the
tools chosen. Tools that are ‘open-source’ and supported by their users are
safer options.
What causes congestion
Data networks form the backbone of any business, whether it is a distributed
or centralized computing environment. The network empowers the enterprise by
integrating business functions and enabling effective communication among
various processes. Hence, proper planning, dedicated implementation, post
integration support and training form a significant part of network management.
The biggest challenge is to keep the existing network highly available or
minimizing the impact on the existing networking environment. Every growing
organization faces the problem of congestion as increasing traffic results in
clogging of network pathways.
An increase in number of users on a single network causes delays as all
networks have a limited data-carrying capacity. To add to it, high demand from
networked applications, such as groupware, e-mail with large attached files and
other bandwidth-intensive applications, such as desktop publishing and
multimedia result in congestion. When the load is light, the average time from
when a host submits a packet for transmission until it is actually sent on the
LAN is relatively short. When many users are vying for connections and
communicating, the average delay increases making the network appear slower. In
extreme circumstances, an application can fail completely under a heavy network
load. Sessions may time-out and disconnect, and applications or operating
systems may actually crash. To stay in control, you need tools to track and
monitor all activity that has the potential to affect the network’s
performance.
Performance indicators
Network performance management includes the processes of quantifying,
measuring, reporting, and controlling of responsiveness, availability, and
utilization for the different network components.
The efficiency of any network is determined by the speed at which users can
perform their tasks. In order to minimize backbone congestion and ensure smooth
performance, one has to carefully design the network. In a properly designed
small to medium-sized network environment, 80% of the traffic on a given network
segment is local (destined for a target in the same workgroup). And not more
than 20% of the network traffic should need to move across a backbone (one that
connects various segments or ‘subnetworks’).
Backbone congestion can indicate that traffic patterns are not meeting the
80-20 rule. In this case, rather than adding switches or upgrading hubs, it may
be easier to improve network performance by moving resources (applications,
software programs, and files from one server to another, for example) to contain
traffic locally within a workgroup. You could also relocate users (logically, if
not physically) so that the workgroups more closely reflect the actual traffic
patterns or you could even add servers so that users can access them locally
without having to cross the backbone.
Ethernet switches, Fast Ethernet hubs, and Fast Ethernet switches immediately
and dramatically improve network performance compared to traditional shared
10-Mbps hubs in a heavily loaded network. Adding these devices to your network
is like adding lanes to a highway (in the case of a switch), increasing the
speed limit (in the case of a Fast Ethernet hub), or both adding a lane and
increasing the speed limit (in the case of a Fast Ethernet switch).
Cabling is an additional equipment consideration when deciding on switched or
Fast Ethernet. Switched Ethernet runs on the common two-pair Category 3 cabling
that many companies have installed as well as Category 4 and Category 5 UTP
(unshielded twisted pair) cabling.
The outsourcing option
Once the equipment lands at the customer premises, the network needs to
perform to pre-defined performance levels, within extremely short times after
delivery, since the cost clock starts upon delivery. It is therefore imperative
that the system integrator be optimally skilled and responsive to meet your
support needs and optimize network costs. Ideally, one should avoid multiple
vendors for their system requirements but in unavoidable cases, there should at
least be proper plans for integration.
Most service providers are enhancing their offerings to include network
related managed services. The performance then depends on delivery by service
providers and tightly defined and measured service level agreements (SLAs). The
SLA contract defines minimum and maximum levels of performance, reliability,
security and cost. Typically, it would include parameters such as network
uptime, application availability, and network and application response time.
In the absence of stable vendor services, enterprises will need to factor in
instances of unplanned downtime, which no management will accept as they fail to
justify the investments in networked operations. Over-engineered, over-priced
network infrastructures should be also avoided to curtail unnecessary costs.
With proper capacity planning and optimum utilization of resources, IT managers
can manage network performance in an effective manner.
Shweta Verma in New Delhi