Network Management should be proactive and pre-emptive - Network Magazine India

Home > Cover Story

Network Management should be proactive and pre-emptive
By Brian Pereira

As business dependence on IT infrastructure increases, network management goes beyond troubleshooting and device configuration

"If top management wants to focus more on service alignment and revenue generation, network management solution is the only path to quick implementation and automation of services"

There are two clear trends in Network Management: Efforts to Integrate NM tools and move towards remote management, perhaps through the Web

As businesses get more consumer-centric and operations switch to 24x7 mode, it becomes imperative to have a network with 99 percent plus uptime this is only possible with good network management.

The term Network Management means different things to different people. It could be something as elementary as a network engineer monitoring network traffic with a protocol analyser or something more intensive, like a team of network consultants using the best monitoring and reporting tools to keep an eye on your network, round the clock.

Network administrators will tell you that managing a network is no easy task, although there are good tools available for automating the whole process. Consultants opine that its crux is management at the application layer, now that hardware has become far more reliable.

"Today Network Management is not just about knowing whether your network link is up or down. Users are now more intelligent and CIOs are a lot more demanding. Network Management now takes into account things like bandwidth utilisation, chokes, and setting up different thresholds and alarms," opines Kiran Bhagwanani, Founder Partner & General Manager, HCL Comnet. "Today the practice is predictive/pre-emptive management rather than fault finding or troubleshooting after an event has occurred."

The complexity of today's networks makes management a sophisticated affair. Enterprise network environments now involve multiple media types, multiple protocols and different platforms. They interconnect with public networks via ISPs or telecommunications companies.

"More complex network environments means potential for connectivity and performance problems in networks is high, and the source of problems is often elusive," says Pawan Sharma, Country Manager-Software Group and Developer Relations, IBM India.

For businesses, the dependence on IT has grown today and it is no longer just IT-savvy companies who invest heavily in IT infrastructure. Naturally, top management has high expectations on IT, and IT staff need to ensure constant QoS.

"If top management expectation from IT is to focus more on business service alignment and revenue-generating activities, a network management solution is the only path to quick implementation and automation of services," says Amit Chatterjee, Country Manager-Software Solutions Organisation, Hewlett-Packard India. "The management solution can proactively generate alarms when business service degrades in performance due to systems and networks."

Besides this, there are other reasons why companies find it compelling to go in for a network management solution. These have been summarised as:

Automation: Networks grow as businesses expand. In a heterogeneous environment it becomes imperative to automate network management.
Availability: With the globalisation of operations, businesses need to make their networks available round the clock. Business entities like suppliers, customers, and clients access servers and databases via VPNs or Web interfaces.
Prioritisation: Prioritisation of network traffic to handle mission critical application responses proactively can only be achieved when network management solutions are implemented within the network.
Dependence on IT: Business functions and processes are becoming increasingly dependent on IT infrastructure. Mission critical applications and databases (hosted on servers) are the crown jewels for many businesses. So if the network fails, businesses lose money and their reputation, even customers/business. Proactive network management can minimise failure/downtime.
Staff training: With the deployment of new technologies on enterprise networks, it becomes necessary to retrain skilled IT staff, which is in perennial shortage. A network management solution can shorten the learning cycle. If IT staff is overwhelmed with daily reactive tasks, a network management solution helps.

MANAGEMENT

As we said earlier, network management means different things to different people. An administrator has to ensure optimal performance even at peak times, when traffic is high. Network management broadly involves monitoring network traffic and devices; troubleshooting & diagnosis; systems configuration; regulation of bandwidth & resources; testing systems for tolerance; upgrading & maintaining networks; network security, data protection; ensuring high-availability; design and integration of new systems (such as security or storage systems).

Some of the processes involved in network management include: auto-polling of network devices and high-end workstations; generating real-time graphical views of network topology changes and traffic; data capture and report generation (reports could indicate traffic composition, protocol and application utilisation); correlating and management of SNMP traps (more on protocols later).

In general, network management is a service that employs a variety of tools, applications, and devices to assist human network managers in monitoring and maintaining networks.

MANAGEMENT TOOLS

There is a wide range of Network Management tools available and one should carefully select and implement the right ones. It is important to consider how well a set of tools integrates with each other. When selecting management tools also look carefully at the partnerships vendors have formed with other management software vendors.

One could go in for enterprise-wide network management solutions like HP-OpenView, CA Unicenter, IBM Tivoli, CiscoWorks and others. These solutions handle a range of functions broadly classified under Security, Storage and Enterprise Management. They have several modules that handle specific functions like: availability management, performance management, service level management, data management, network/systems management; storage management, console management, device management, event management, asset management etc.

Then there are highly specialised tools like protocol analysers, port scanners, security scanners, device monitoring/configuration tools, network discovery & mapping software; network monitoring and alerting tools; risk/vulnerability assessment tools (with auditing/reporting) etc. Examples of such tools are Sniffer (Network Associates), Cheops-ng, Cybercop (Network Associates), Nesus, Concorde, WhatsUp (IP Switch).

Network Administrators may also use hardware devices such as probes and network analysers. These are mainly used for checking cable faults, for network trending and frame capture. Notable examples of hardware-based analysers are Network General Sniffer, HP Internet Advisor, Fluke LANMeter, and Trend FastLAN900.

Network analysers are also of the software variety and some examples are: Chevin Personal Protocol Analyser, Cinco's NetXRay, EtherPeek and Novell's LANalyzer.

PROTOCOLS & STANDARDS

Network management software and tools rely heavily on management protocols and standards. A Network Administrator must be well-versed with popular protocols like Simple Network Management Protocol (SNMP) and Common Management Information Protocol (CMIP).

There is a wide spectrum of applications for network management and so integration becomes crucial. SNMP helps integrate all the management tools by sending SNMP traps between them so that all events are collected in one place.

There have been efforts towards creating common standards through Policy Based Networking. A number of companies have announced tools that support the specification and deployment of policies. Much of this work is focused on developing policies for quality of service management within networks. The Internet Engineering Task Force (IETF) and Distributed Management Task Force (DMTF) is actively working on standards related to this area.

The DMTF's Web-Based Enterprise Management (WBEM) is a set of management and Internet standard technologies aimed at unifying the management of enterprise computing environments. WBEM provides the ability for the industry to deliver a well-integrated set of standards-based management tools by leveraging emerging Web technologies. In other words, these standards will enable cross-platform and cross-application access to system information from any management console using Web technologies.

In the same vein, IETF's Policy Framework Working Group is working on an overarching management framework. Data models, common access methods and groupings have been created.

TRENDS

The protocols and standards discussed above are also direct indicators of emerging trends in Network Management. Two clear trends can be observed firstly there are efforts towards integration of network management tools and secondly, everyone is considering remote management, perhaps through the Web. Here are some emerging trends:

Management Services Providers: This new breed of service providers is making heavy investments in Network Operations Centres (NOC). With state-of-the art equipment and teams of highly skilled technicians, MSPs will manage an organisation's network remotely, round the clock. The MSP approach can free organisations from the hassles of investing in new network infrastructure every few months, and from employing skilled IT staff, which is itself in short supply. In India, HCL Comnet, Bangalore Labs, WiproNet and Convergent Communications have NOCs and offer Managed Services. Other ISPs will soon offer such services.
Integration: Initiatives and standards from IETF and DMTF will bring about better integration between network management products. Tighter integration will result in improved user interfaces/common views for managing several products. But later, we may also see integration at the database level, wherein management data from disparate resources is intelligently combined.
Management portals: xSPs will offer value-added services like network management via customised portals. Through these portals IT staff will have customised views of network data by pulling together displays from disparate management products.

Brian Pereira can be reached at brianp@rediffmail.com

Top 10 Rules for Network Management

1. Knowledge of network - to be able to discover network elements quickly and efficiently track new elements that are periodically introduced on the network.

2. Capacity planning - to meet the increasing demands of business from IT infrastructure. Planning to provision capacity before it expires requires proactive tool based solutions.

3. SLA monitoring - being able to monitor the service levels within a network so that performance of various elements within the network can be managed.

4. The-user-knows-first - reactive management of networks and isolating problems before they impact the end users, is a key issue to be addressed. Only then is management proactive.

5. Dynamic resource allocation - on the fly allocation of resources to meet fluctuations in business with IT infrastructure needs prioritising the network, application and traffic on a dynamic basis. This is difficult but crucial in the current context.

6. Heterogeneity breeds complexity - rapidly growing network with heterogeneous platforms creates issues of interoperability and connectivity. Deploying a solution to manage this is a big challenge.

7. Event correlation - if every event generates an alarm, depending on implementation, there could be situations where thousands of alarms are generated within a very short time span. It is therefore important to be able to correlate all alarms to certain IS Services going down, or certain capacity augmentation efforts. So the biggest challenge is to be able to correctly correlate and generate business views for events as they occur.

8. Determining the bottlenecks - simple implementation and fast discovery of network elements will lead to quicker isolation of bottlenecks as and when service levels are violated. Network managers must have the ability to quickly isolate a bottleneck, take remedial action and contain a potential disaster or downtime on a business critical network.

9. Manage bandwidth more effectively - the exponential growth of data traffic has contributed significantly to current loads on corporate and service provider networks. With voice and video applications driving network demand, bandwidth requirements continue to grow at a rapid rate. Meanwhile, mission-critical traffic must battle against both high frequency & high-bandwidth competitors, since the IP network itself provides nothing but best-effort service to all traffic types.

10. Quick problem diagnosis - data trending, data log analysis and data collection is a laborious, time consuming & repetitive work. Easy data collection and trend analysis through out-of-the-box reporting will help solve this issue for network managers.

Courtesy: Hewlett-Packard India Ltd.