STREAMING VIDEO OVER TFRC WITH LINEAR THROUGHPUT EQUATION***

__________________________________________

* Kielce University of Technology.

** AGH University of Technology.

*** This work is supported by Polish Government under Grant No. N517 012 32/2108 (years 2007-2009).

Agnieszka CHODOREK*

Robert R. CHODOREK**

STREAMING VIDEO OVER TFRC WITH LINEAR THROUGHPUT EQUATION***

The TCP-Friendly Rate Control (TFRC) protocol manifests strong equality towards competing TCP or TCP-friendly flows. Although the RFC 3448 suggest, that TFRC is suitable for multimedia, this equality is a great disadvantage in the case of transmission of multimedia over the TFRC.

The TFRC emulates TCP-like congestion control using the TCP throughput equation. In the paper, we substitute TCP throughput equation recommended for the TFRC by linear throughput equation. Simulation results show, that proposed solution is more suitable for multimedia than the equation proposed in RFC 3448. Experiments were carried out using an event-driven ns-2 simulator, developed in U. C. Berkeley.

Keywords: TCP-friendly protocol, congestion control, multimedia

1. INTRODUCTION

Phenomenon of collapse of TCP transmissions, which competes for bandwidth with multimedia over RTP/UDP or UDP, was the base to design so-called TCP- friendly transport protocols. One of the most known and the first standardized TCP-friendly protocol was the TCP-Friendly Rate Control (TFRC) [1]. This multipurpose protocol was designed to carry different kind of data, including real- time multimedia.

TCP-friendly transport protocols implements TCP-like congestion control and behaves under congestion like the TCP. Among others, the equally share throughput of bottleneck link with TCP flows or other TCP-friendly flows. This feature is a great advantage in the case of bulk data transfer, because it allows to achieve Quality of Service (QoS) appropriate for each transmission. In the case of real-time multimedia transmission we can see opposite tendency. If flow equality is contrary to real-time requirements, we observe degradation of QoS of multimedia

transmission. The deeper a conflict between the equality and real-time becomes, the lager degradation is observed [2][3].

The TFRC emulates TCP-like congestion control using the TCP throughput equation. The equation is used for estimation of instantaneous throughput of TFRC under congestion. In the paper, we substitute TCP throughput equation recommended for the TFRC by linear function of packet error rate. The aim of such substitution is to develop a transport protocol, which is more suitable for multimedia than the TFRC and more TCP-friendly than the RTP.

The paper is organised as follows. Section 2 briefly describes the TFRC protocol. Section 3 proposes a linear function, which will be used as a throughput equation for the TFRC. Section 4 describes simulation experiments. Section 5 presents simulation results of TFRC and TCP transmissions in shared link. Section 6 summarizes our experiences.

2. THE TFRC PROTOCOL

The TFRC protocol presents a modern approach to transport layers protocols, which tread protocols as a set of building blocks – independent components, from which transport protocols are assembled [4]. The TFRC is a congestion control building block, which was designed to be reasonably fair when competing for bandwidth with TCP flows. As other control systems, the TFRC consists of:

– a controller, which takes decisions about value of controlled quantity,

– a control device, which adjusts controlled quantity to the value given by the controller.

In the case of the TFRC, controller (the congestion control mechanism) evaluates output throughput of the flow using so-called TCP throughput equation, which is, in fact, an analytical model of the TCP behaviour under congestion. The equation describes TCP throughput as a function of packet error rate. The TFRC uses Padhye’s model of TCP throughput, described in [5][6]. According to this model, throughput of the TCP protocol (and, in result, the TFRC throughput) is equal to:

(

)

8 12 3

3 ) 2

(

PER PER

PER PER

C RTT

PER MSS T

⋅ +

⋅

⋅

⋅

⋅ +

⋅

= (1)

where PER denotes a packet error rate, T is a TCP throughput, and C is scale coefficient.

An output throughput of the TFRC is adjusted to the value given by the controller using rate control mechanism. This mechanism vary TFRC sending rate in packets per second.

Authors of the RFC 3448 recommend the TFRC as suitable for applications such as telephony or streaming media. They suggest also, that the TFRC could be used in a transport protocol such as Real-time Transport Protocol (RTP) [7], which is commonly used as a transport protocol for audio and video transmission.

3. LINEAR THROUGHPUT EQUATION

Typical for TCP-friendly protocols equality towards competing TCP flows causes, that sometimes the TFRC isn’t able to meet real-time requirements of multimedia transmission. It means, that TCP is too aggressive (when compared with TFRC) to allow the TFRC to preserve QoS for multimedia traffic.

The protocol, which is aggressive enough to force QoS of carrying multimedia traffic is the RTP. Because RTP implements neither congestion control, nor flow control, nor error control, offered traffic will be reduced only by packet losses. As a result, in a network that is well-dimensioned for multimedia, analytical model of RTP throughput should depend only on target bit rate of carried multimedia stream and packet error rate. Thus, RTP throughput equation should be as follows:

0 0

)

( t

S t PER S

T = ^p − ^l (2)

where PER denotes a packet error rate, T is a RTP throughput, S_p is an amount of information (in bits) send in RTP packets (both in headers and payloads) during the time t₀, S_l is an amount of information carried in RTP packets, which were lost or damaged during the time t₀, t₀ is an observation time.

Note, that above analytical model of RTP throughput describes both transmission of streaming media over RTP/UDP and transmission of streaming media over UDP.

In the paper, we propose to substitute the TCP throughput equation (1), used by the TFRC, by linear throughput equation:

^T(^PER)=^TBR

(

)

Because

TBR t

S_p

=

0

(4) and

PER S

S

p

l = (5)

the linear throughput equation describes, in fact, the RTP/UDP throughput as a function of packet error rate.

Because proposed equation bases on the RTP mode, we believe it is aggressive enough to preserve real-time character of transmitted steaming media. However, it doesn’t mean that TFRC will behave under congestion like the RTP, if linear throughput equation is used. The RTP protocol doesn’t implement congestion control. It is not able to change transmission rate due to congestion.

The TFRC have congestion control, although usage of linear throughput equation causes, that it is congestion control in “light” version. Sending rate is reduces only by packets, which are lost due to congestion. It means, that TFRC can’t aggressive avoid congestions, but it don’t allow the congestion to grow up.

4. SIMULATION EXPERIMENTS

Simulation experiments were carried out using single-bottleneck topology (Fig. 1.). Senders S are connected to the router R1 via 100 Mb/s links with 1 µs propagation delay. The same links are used to connect receivers R and router R2.

Routers are connected via 4 Mb/s bottleneck link with 10 ms propagation delay.

R1 R2

S R

S1TCP

R1 TCP

SN TCP

RN TCP

... ...

100 Mb/s 1 µs

100 Mb/s 1 µs 4 Mb/s

10 ms

Fig. 1. Topology of simulated network.

Constant Bit Rate (CBR) video stream is transmitted between S and R end- systems and target bit rate of the stream is equal to B. Because we assume, that network is well-dimensioned for multimedia, 0 Mb/s ≤ B ≤ 4 Mb/s. Real-time CBR transmission is carried out using the TFRC and modified TFRC with linear throughput equation. For the sake of comparison, RTP/UDP protocols also are used. FTP over TCP transmissions are carried out between the pair of nodes S_i^TCP and RiTCP

, i = 1,…,N. All transport protocols, used in experiments, have the same size of data packets – 1000 B (960 B of data + 40 B overheads).

During experiments we investigate achieved throughput (both for multimedia and bulk data transfer. Experiments were carried out using Berkeley’s ns-2 simulator [8].

5. SIMULATION RESULTS

In the first experiment we changed the number of competing TCP flows N from 0 to 10. Target bit rate of CBR transmission was set to 1 Mb/s (1/4 of throughput of the bottleneck link). Results are shown in Fig. 2.

0,00E+00 2,00E+05 4,00E+05 6,00E+05 8,00E+05 1,00E+06 1,20E+06

0 1 2 3 4 5 6 7 8 9 10

No. of TCP flows N

Throughput [b/s] .

RTP modified TFRC TFRC

Fig. 2. Throughput of CBR transmission as a function of N.

Simulation results show, that CBR video transmission will preserve their real- time character if modified TFRC which linear equation is used in transport layer.

Streaming video over classic TFRC (with TCP throughput equation) causes strong degradation of CBR connection in the case of larger values of N.

Usage of proposed solution instead of classic TFRC allows to achieve throughput of CBR stream comparable with throughput of CBR over RTP.

Moreover, parameters of the TCP transmissions are approximately these same, as observed when classic TFRC is used. It means, that linear equation avoids collapse of TCP connections and allows the TCP to utilize available bandwidth (bandwidth of the bottleneck link reduced by target bit rate of multimedia stream).

In the second experiment we changed throughput of CBR transmission B from 0.5 Mb/s to 4 Mb/s (a throughput of the bottleneck link). The number of competing TCP flows was set to 1. Results of experiments are shown in Fig. 3.

0,00E+00 5,00E+05 1,00E+06 1,50E+06 2,00E+06 2,50E+06 3,00E+06 3,50E+06 4,00E+06 4,50E+06

0,5 1 1,5 2 2,5 3 3,5 4

Target bit rate [Mb/s]

Throughput [b/s] .

RTP modif ied TFRC TFRC

Fig. 3. Throughput of CBR transmission as a function of B.

As we can see in Fig. 3., TFRC with linear equation allows to transmit real- time multimedia even if target bit rate of CBR stream is close to throughput of bottleneck link. Both RTP and classic TFRC were able to carry out real-time transmission up to about a half of throughput of bottleneck link (at least in this experiment). In the case of both modified TFRC and classic TFRC, concurrent TCP stream was able to utilize all remaining bandwidth of bottleneck link.

5. CONCLUSIONS

Although authors of TFRC suggests, that the protocol is suitable for multimedia transmission, it isn’t enough aggressive to meet QoS requirements of carried streaming media when competes for bandwidth with the TCP. In the paper we propose to substitute original TFRC throughput equation by linear throughput equation. This substitution makes the TFRC more aggressive, what allow the protocol to preserve real-time character of transmitted flow not worst than the RTP or the UDP protocol. Moreover, in situations, where usage of the RTP causes collapse of TCP transmissions (or, at least, worsening of the QoS of one or more TCP flows) the proposed solution is enough “friendly” for competing TCP flows that they are able to equally share remaining bandwidth. Such results allows us to believe, that proposed linear equation is more suitable for multimedia transmission than the equation originally included in the RFC 3448.

REFERENCES

[1] Handley M., Floyd S., Padhye J., Widmer J.: TCP Friendly Rate Control (TFRC):

Protocol Specification. RFC 3448. January 2003.

[2] Chodorek A., Chodorek R.R.: Applicability of TCP-friendly protocols for real-time multimedia transmission. Submitted to PWT’2007, Poznań, 2007.

[3] Chodorek A.: Streaming video with TFRC - simulation approach. Proceedings of SympoTIC’04, 24-26 Oct. 2004.

[4] Chodorek A., Chodorek R.R., Pach A.R.: Dystrybucja danych w sieci Internet.

WKŁ, Warszawa 2007.

[5] Padhye J.: Model-based approach to TCP-friendly congestion control. Department of Computer Science, University of Massachusetts at Amherst, 2000 (Ph.D. Thesis).

[6] Padhye J., Firoiu V., Towsley D., Kurose J.: Modeling TCP Throughput: A Simple Model and its Empirical Validation. Proceedings of ACM SIGCOMM 1998.

[7] Schulzrinne H., Casner S., Frederick R., Jacobson V.: RTP: A Transport Protocol for Real-Time Applications. RFC 3550. July 2003.

[8] Fall K., Vradhan K.: The ns Manual. http://www.isi.edu/nsnam/doc/ns_doc.pdf