DOC PREVIEW
UCLA COMSCI 218 - TCP Westwood and Bandwidth Estimation

This preview shows page 1-2-3-4-5 out of 16 pages.

Save
View full document
View full document
Premium Document
Do you want full access? Go Premium and unlock all 16 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 16 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 16 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 16 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 16 pages.
Access to all documents
Download any document
Ad free experience
Premium Document
Do you want full access? Go Premium and unlock all 16 pages.
Access to all documents
Download any document
Ad free experience

Unformatted text preview:

BATCP flowsComputer Science DepartmentUCLACS 218 Project1. Introduction2. Related works3. Rationale of the idea4. Simulation environmentFig. 5.1 – Single TCP flow on a bottleneck link of 5Mb, with no errors6. Conclusions.TCP Westwood and BandwidthEstimation with Sender-Side InformationAUTHOR: Claudio Enrico Palazzi - ([email protected])TUTOR: Dr. Giovanni Pau Computer Science DepartmentUCLACS 218 ProjectFall 20031. IntroductionEven if, at present, packet switched technology can be considered mature, new applications as peer-to-peer, remote working, video/audio streaming, television distribution and interactive onlinegaming present new challenges that require at least an adaptation or, in some cases, a completeredesign of the currently used mechanisms. New applications usually necessitate of capabilities thatwhere not included in the original design of the Internet, thus asking for solutions able to providethe efficiency needed in order make the new products widely adopted by the final users. In order to really understand the complexity of this scenario we also have to mention the highnumber of people using wireless technology. Cellular phones, laptops and PDAs are devices ownedby a very elevated and still increasing percentage of people and, with the introduction of the higherdata rates of the third generation (3G) of mobile systems [10], we can easily foresee a raisingrequest for new services able to push people in purchasing this new technology. The new “killerapplications” able to alter the trajectory of the market will probably come from the interactionbetween the new devices and the Internet. Indeed, virtual libraries, video-telephony over IP,videoconferencing, games, remote-medicine, video and music on demand and locality basedinformation are only few of the innumerable services that will be available in every place and atevery time.On the other hand, the interaction between mobile devices and the Internet, coupled with the needfor an efficient and reliable data transfer have generated several unresolved problems. The mainchallenges that designers are called to face in a mobile scenario can be summarized as follows [5]:- Limited and variable bandwidth;- High and variable latency that could be independent from the network congestion level;- High BER (Bit Error Rate) and consequently losses of many packets;- Frequent and also extended disconnections, depending on the user’s mobility;- Limited computational and energy resources of the mobile device;- Mobility of the connection (horizontal and vertical handoffs).The wireless environment especially affects applications that rely on constant (or slowly changing)channel conditions or on a reliable transmission protocol as TCP. The former case include, forinstance, video and audio streaming: the quality of the stream, the compression level, the deliveryrate and the anticipated buffering strongly depend on the detected channel conditions such asavailable bandwidth and expected delay. In order to work properly, the system should be able todetect in time changes and react promptly adjusting the parameters of the stream. Focusing on the TCP (Transmission Control Protocol), indeed the most popular transport protocolfor reliable data delivery, we have also to notice that it was designed in a time when networks wereexclusively based on wired technology. For this reason, the flow control and the congestion controlfunctions fail when introduced into a wireless context [13]. To really understand the reasons of thisfailure we should remember that traditional TCP uses packet losses as a metric to evaluate thecongestion level of the network. Consequently, when a packet is considered to be lost, TCP reducesthe data sending rate. In presence of numerous losses related to non-congestion factors, as in awireless environment, this behavior is not appropriate and causes a consistent underutilization of theavailable bandwidth on the link [4]. As a result, we experience a considerable decline in themeasured performance compared to the quality of service expected by the users. Having a means toprecisely estimate the accessible bandwidth could help in discriminating between the congestionrelated losses and wireless error losses. In order to avoid potential overestimation of the availablebandwidth it could be useful to have a tool able to provide a capacity measure of the bottleneck. Notonly this value could be used as an upper bound for data sending rates, but it could also be put ingood use in other applications as, for instance, to determine appropriate routes/trees for multicastoverlay networks [12]. In Section 2 we briefly present related works which regards estimation either of bandwidth orcapacity, in Section 3 we propose the reasoning that brought us to design our new link capacityestimator, in Section 4 we explain the simulated environment used to test the new mechanism, inSection 5 we show the results obtained in simulations by our scheme and Section 6 concludes thiswork.42. Related worksIn recent years, many researchers have focused their studies on the obstacles present in a wirelessenvironment proposing various alternative techniques [14]. Several solutions proposed to face thoseproblems relies on the ability to estimate the factual capacity or the available bandwidth. Equippedwith these information, protocols could be able to set the most appropriate sending parametersdecoupled from the losses. Tsaoussidis and Badr propose the use of special pairs of probe packets after each loss [17]. Whenone of these couples reaches the receiver, the measured delay of these probing packets, is used tounderstand if the network is congested and, only in this case, to diminish the data sending rate. Ifthe wireless link is experiencing a disconnection or a fading, the probe cycle is extended, thusavoiding further data losses and consequently erroneous restrictions of the sending rate. Theirpurpose is also obtaining a better energy resource consumption. This protocol requires modificationat both sender and receiver side in order to handle with probing packets: this lack of compatibilitywith the current implementation of the TCP seriously affects its real development possibility.A new end-to-end transport protocol is also proposed by Sinha et al. [18]. They face the highnumber of errors and the


View Full Document

UCLA COMSCI 218 - TCP Westwood and Bandwidth Estimation

Documents in this Course
GSM

GSM

59 pages

Chord

Chord

30 pages

10_2

10_2

9 pages

13_4

13_4

10 pages

RAP

RAP

17 pages

46_4

46_4

9 pages

32_4

32_4

10 pages

umts

umts

39 pages

AdHoc-MAC

AdHoc-MAC

29 pages

rma

rma

8 pages

Lecture

Lecture

29 pages

Load more
Download TCP Westwood and Bandwidth Estimation
Our administrator received your request to download this document. We will send you the file to your email shortly.
Loading Unlocking...
Login

Join to view TCP Westwood and Bandwidth Estimation and access 3M+ class-specific study document.

or
We will never post anything without your permission.
Don't have an account?
Sign Up

Join to view TCP Westwood and Bandwidth Estimation 2 2 and access 3M+ class-specific study document.

or

By creating an account you agree to our Privacy Policy and Terms Of Use

Already a member?