1 / 7

Quality of Service in Internetworks

Quality of Service in Internetworks. Exercise 1 Rate control. Example setting. A topology consisting of: Customer site with traffic sources/sinks Access Provider Edge device linked to CE with sufficient delay Arbitrary headend of traffic. Token bucket rate control.

vondra
Download Presentation

Quality of Service in Internetworks

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript

  1. Quality of Service in Internetworks Exercise 1 Rate control

  2. Example setting • A topology consisting of: • Customer site with traffic sources/sinks • Access Provider Edge device linked to CE with sufficient delay • Arbitrary headend of traffic

  3. Token bucket rate control • Decide on three or more distinctively different unicast flow types • UDP vs TCP • Long vs short-lived • Bandwidth consumption (and possible variation thereof) • Burstness of outbound traffic • Simulate said flow types on given topology • Create different token bucket policies and install them one at a time on PE ingress port • Limit the total accepted throughput to 1, 2, and 10 Mbit/s • Also remember to add some delay to the link • Send more traffic on the link that is accepted at PE

  4. Time sliding window rate control • Create a group of time sliding window policies • Perform the simulations in previous manner by setting the policies into PE ingress port and flooding the link • You can use same flow types as in token bucket but you can also create new ones in addition • How does the window size affect traffic behavior? • Does one type of traffic benefit more from sliding window than another?

  5. For both mechanisms • How do aggregate throughput, delay and other traffic properties behave with different policies? • How does packet loss occur over time • for TCP applications? • for UDP applications? • How much goodput is attained by an individual flow when tens or hundreds of flows are aggregated into the link? • Also for TCP/UDP • How do the methods differ from each other? • Can you say explicitly that one is better than the other? • Can you claim that either is better for a specific flow type?

  6. Weapon of choice • You can use whatever tool you see fit to perform the simulation • Omnet • ns2, ns3 • Matlab • Perl + awk + lots of glue • That extra Cisco 1200 you have in your dorm room • Virtual machines... • However, remember that this can distort the results in the worst case scenario • We’re interested in your reasonings based on the simulations • But feel free to add the simulation code as an appendix if you want to

  7. Deliverables • Do the aforementioned tasks, and return a written report of 2-3 pages • Emphasis on own observations, not producing textbook answers about mechanisms • Include any other thoughts about rate control that occurred to you during the exercise • This is not an English grammar course! • Good textual output is only appreciated, not seen as a heavy grading factor • Return date: 20.4.2009

More Related