DistributedNetworks DistributedNetworks


DNS Lookup  «Prev 

Bandwidth Utilization

You can estimate bandwidth utilization by using existing measurements obtained through applications such as the Microsoft Network Monitor and the Windows 2000 System Monitor. Broadcast or collision domain size is determined by the number of hosts on the network that are on the same side of a layer 2 device.
You can evaluate the routing configuration of a network by reading the existing network documentation, or by creating your own documentation. Distance Vector Delays are related to the number of hops data must travel from source to destination, and any delays that take place because of link or router saturation.
You can estimate application data flow requirements by calculating the average packet size created by the application and then multiplying that by the number of packets per unit time.
Bandwidth utilization is a network statistic and there are a couple factors to take into account when monitoring bandwidth utilization.

First Factor

The first factor is understanding the difference between 1) bandwidth utilization statistic from a single segment and 2) an aggregate of segments. Administrators often want to know, or are asked the question,
"What is the bandwidth utilization of our network?"
The problem is that even on the most basic of networks, this is almost an impossible statistic to determine.

This is because that even on a flat network, there are typically multiple segments, sites, and links that cannot realistically be totaled. For example, if you have 10 segments that you have gathered bandwidth utilization information on for a number of weeks, averaging that data will result in a meaningless number that does not reflect any real statistic. For instance, if nine of the segments have an average utilization of 10 percent, and one segment reaches 95 percent utilization, then "Network Utilization" would be 18.5 percent. This number is of no value because it hides the problem of the single site and increases the utilization of other sites. However, if each segment is monitored separately, the administrator will be able to stay on top of the network, identifying and resolving issues before they even become a problem. Proactive utilization management can save any company significant time and money.

Second Factor

Another factor to take into account when monitoring utilization is the source of bandwidth consumption. VoIP, database applications, and graphic design programs are all examples of common applications that are known to consume fairly large amounts of network bandwidth. Often, when considering deployment, an administrator feels obligated to increase available bandwidth so that the application will not experience issues. However, the current budget may be tight and purchasing additional bandwidth may require solid justification. The first step in justifying additional bandwidth is to determine what the current utilization level is on a network segment or link, and second, to trend the actual requirements of the application in a test lab. Simulating usage loads and comparing that information to actual findings to determine if an increase in available bandwidth will be required. Those results may show that it is pertinent to increase bandwidth capacity in order to keep the applications and network running smoothly, and you will have the data to prove it. On the other hand, you might see a negligible increase in utilization, saving the time and money that would otherwise be spent on an unnecessary upgrade.

DHCP only sends about 4 packets of data so its a tiny amount. We have had sites with more than 100 machines running on the end of a 4Mbit DSL connection with DMVPN over the top, as well as domain traffic, DNS, HTTP etc (no server onsite at all) We never saw any issues, even when rebooting all the machines at 4AM every morning. Stick your DHCP server on a dedicated port, monitor said port with something like PRTG \ MRTG over a 24 hour period and count the packets and bits per second.