Bandwith Utilization Policy

Overview

Network connectivity and bandwidth are essential resources for members of the Rose-Hulman community. The internal campus network encompasses routers and switches that interconnect the campus buildings using a mixture of fiber optic and copper cables. With the distributed nature of the internal network, i.e. people connect to the network using wired and wireless interfaces in many campus locations, the internal network generally has sufficient bandwidth for the needs of the campus community, e.g. email, learning management, etc.

However, Rose-Hulman's Internet connection, currently 300 megabits per second (Mbps), is a single, limited resource that everyone accessing the Internet must share. As such, excessive utilization by a small number of individuals can significantly impact the ability of others to effectively access and use the Internet. The Rose-Hulman network and Internet service are provided to the entire community to support the Rose-Hulman mission, and automated systems are employed to limit disproportionate use of the campus Internet connection. The Institute expects all members of the community to use these resources in a responsible manner, respecting the rights, needs and privacy of other persons utilizing the network.

The bandwidth utilization policy applies to individuals, not devices, e.g. desktop and laptop computers, smart phones, game consoles, etc. This means that the aggregate utilization of all devices registered to an individual will count toward the utilization thresholds specified below. Internet utilization will be monitored on a continuous basis and changes may be made to the thresholds listed below to maintain an effective service level.

Details

There were several goals when designing this bandwidth utilization policy. They were:

• Minimize the impact for typical use cases, and provide adequate bandwidth for web surfing at the lowest step (see table below).
• Target an average network utilization of 85% on the campus Internet connection.
• Craft a policy that would encourage network utilization during off-hours, i.e. 2:00am until 6:00am, when network utilization typically drops to its lowest levels.

The three steps, along with the corresponding bandwidth rates, are shown in the table below. The thresholds are measured in a 36-hour sliding window. For example, there are no bandwidth rate limits imposed on an individual until the total amount of data transferred across all registered devices reaches 4 gigabytes. At that time, the individual will have a 1024 kilobit per second (kbps) bandwidth rate limit. When a total of 4.5 gigabytes of data is transferred, the individual's bandwidth rate limit is set to 160 kbps. Since the amount transferred is measured in a 36-hour sliding window, reducing Internet transfers will allow the bandwidth rate to recover, eventually to the unrestricted rate. Note that download and upload transfers are counted separately, and exceeding the thresholds listed below in either direction will result in the stated bandwidth rate.

For example, when these limits are applied over 72-hours of constant Internet downloading, approximately 5 gigabytes of data will be download during the first 24 hours and approximately 9.5 gigabytes of data will be download during the entire 72 hour period. Compare this with the previous policy, which specified a 3 gigabyte transfer limit in 24 hours or 5 gigabytes in 72 hours. In summary, there are no email notifications, the transfer limits are higher, and bandwidth rates will recover over time if Internet usage is reduced.

As mentioned above one of the design goals was to push some of the large downloads into the off-hours, i.e. from 2:00am through 6:00am. Data transfers from the Internet during these hours will only be counted at 75%. For example, a 1 gigabyte download during the off-hours will only count as 750 megabytes against an individual's transfer threshold. The off-hours "discount" will also be enabled on weekends beginning at 6:00pm on Friday and ending at 6:00am on Monday and all day during academic breaks, e.g. Fall, Spring and quarter breaks.

Internet performance will be monitored for average and total utilization, and changes to the thresholds, bandwidth rates, and off-hours adjustment may be made to maintain effective service levels for the Institute.

Peer-to-Peer Bandwidth Limits

A Peer-to-Peer, or P2P, architecture is based on the concept of distributed and individually managed computers cooperating on an ad hoc basis to share resources like files, computation cycles, or network bandwidth. There is no central authority managing or coordinating the resource sharing. P2P protocols can be a highly efficient and reliable mechanism for resource sharing.

File sharing is a very popular application of the P2P architecture, and there are a number of programs that facilitate this function. This easy access to content and resources on the Internet can saturate, i.e. fully consume, the campus Internet connection. In addition, P2P applications like Bit Torrent, Direct Connect, and eMule make it very easy for an individual to consume a disproportionate amount of bandwidth. The bandwidth rate limits for network traffic classified as P2P are listed in the table below.