Analyze Network Performance

Preface

This paper intends to describe the needs of detecting bottleneck of networks, and consequently increase ROI to achieve the business objective. The topic is mainly focused on performance of network.

Challenge

The efficiency of IT system will directly affect the productivity and consequently influence the profit.

Unlike the traditional industry which generates revenue by large-scale equipments and manufacture productive factories, the modern enterprises produce revenue most in logical ways and make profits by cost effective Information system.

Today's industries rely highly on information technologies and in a certain level; their productivity can be extremely affected by the efficiency of the IT system.

FREE TRIAL DOWNLOAD

• Efficient network - Swift information flow brings well communication and accelerates information exchange, decision making, and agreement dealing, thus increase the productivity. While the expense of the IT system is fixed, the productivity and revenue is increased, you make more profits.
• Inefficient network - packet delay results in information flow latency, and this would slow down the entire operation of any department which is more or less relied on IT system and then consequently reduce the revenue through low productivity.

As we all know, network is the artery of IT systems, it is the key point to the health of the entire IT system.

Why network performance declined?

Along with the information technology evolves, different new types of applications are introduced to enterprises, organizations. These new app like online webcast, real-time stream media, VoIP, brought added value to network, but as the same time they occupied lots of traffic and bring huge burden to the IT infrastructure. The IT managers worried about whether their network can still function properly when there are too many applications overloaded the infrastructure.

Another reason is abuse of network resource. Bandwidth shortage can be caused by device failure, ISP issue, unexpected non human elements, and, sometimes human behaviors, like p2p download programs: bit torrent, emule etc. Such sort of abuse always occupies a huge bandwidth and cause congestion to other users in a network without QoS policy.

There are also other causes of declined network performance like too many departments, dispersed staff location, flawed network design.

Need to monitor the performance and have your baseline regarding the network

Since too many burdens drag down the network performance and negatively affected the productivity, there must be something to do to fix it.

What is needed is to monitor the network performance.

• Be aware of whether the infrastructure provided adequate support for applications and users.
• Create a baseline indicating in what level the network can function properly without performance decrease.
• Know the crucial business traffic/transactions running in your network.

The procedure to amend the problem is like this:

• When there's a symptom of performance decrease in the network, monitor the transaction which encountered congestion.
• Pinpoint where the bottleneck is based on the statistics from the monitor tools.
• Analyze what the traffics are.
• Determining whether the performance decline is caused by abuse of network resource, or crucial applications have overloaded the infrastructure.
• Based on step 4, implement controls to solve the problem. Either add/replace network device like interface card/switch/router to strengthen the infrastructure, or ban the abuse of resource.

Now what we need is a tool for monitor (preferable with statistics provided), and a tool for analyze traffic.

Solution

Working as a probe, through its Mobility and Accessibility, Colasoft Capsa endows administrators the ability to penetrate into each corner or angle which is not accessible by those large network management solutions. There are always some tiny corners the dust collector can't reach, and the Capsa is right the small brush for you.

Monitor

By way of SPAN build in manageable switches, Colasoft Capsa will accept the network traffic forwarded by the SPAN configuration and account the traffic load, display all kinds of parameters like packet numbers, total bits/bytes, bps, and so on. And a visualized graph can be generated to show the statistics in a vivid way. You will be able to have a baseline in mind based on these statistics, and when the congestion arises, it is easy to pinpoint where the bottleneck is.

Analyze

Only after a clear understanding of what caused the congestion, can you decide what to do next: you shall not just simply cut the traffic in case it is a remote online meeting, nor blindly increase the bandwidth because if the congestion is cause by network abuse, you are then just wasting money. Any consequent operation should base on an analysis of what the traffic is. Colasoft Capsa is a network analyzer which does not only give you a broad view of monitoring, but also an insight view in the depth of frame level. You can have a brief idea of what the traffic are based on the protocol identity.