Dynamic throughput optimization in SCTP

Most current web application use HTTP (Hyper Text Transfer Protocol) and TCP (Transmission Control Protocol) to retrieve objects from the Internet. SCTP (Stream Control Transmission Protocol) is recently proposed transport protocol with congestion control mechanism similar to that of TCP. Waiting delay is an important performance criterion for when transferring web object over the Internet. In this paper, we present an analytical model of mean waiting delay for object transfers over the Internet in a wireless using the SCTP as the transport protocol and compare with the mean waiting in using TCP. Validation of the model using experimental results show that the mean waiting delay for HTTP over SCTP is less than that for HTTP over TCP. This is caused by the small slow-start time of SCTP.

현재 인터넷에서는 객체를 전송하기 위해 HTTP over TCP가 사용된다. SCTP는 TCP의 문제점을 보완하기 위해 제안된 프로토콜로 개별적인 스트림을 이용하여 데이터를 전송하므로 웹 객체의 평균 응답 시간을 감소시킬 수 있다. 본 논문에서는 HTTP over SCTP의 평균 응답 시간을 추정하는 해석적 모델을 개발함으로써 TCP에 대한 SCTP의 효과를 추정하고자 한다. TCP 지연을 구하기 위한 기존의 해석적 모델이 주로 유선 환경을 가정한데 비해, 본 연구에서 제안하는 모델은 다중 패킷 손실이 발생하고, 빠른 재전송이 가능하지 않은 무선 환경을 가정한다. 추정된 평균 응답 시간은 실제 웹 응용을 개발할 때, 종단 사용자의 서비스 품질을 만족시키기 위한 사전 벤치마크로 사용될 수 있다. 추정 모델의 타당성을 검증하기 위해 테스트베드 상에서 성능 실험을 수행한 결과, 모델과 실험결과의 차이가 평균 6% 이내로 매우 작으며, 동시에 HTTP over SCTP의 평균 전송 시간이 HTTP over TCP의 평균 전송 시간보다 작은 것으로 확인되었다. Hyper text transfer protocol (HTTP) over transmission control protocol (TCP) is currently used to transfer objects in the Internet. Stream control transmission protocol (SCTP), an alternative to TCP, which allows for independent delivery among streams, and can thus reduce the mean response delay of web object. We present an analytical model to find the mean response delay for HTTP over SCTP, therefore, estimate the effectiveness of SCTP over TCP. Typical TCP delay models assume the wired environment. On the contrary, the proposed model in this paper assumes the multiple packet losses and wireless environment where fast retransmission is not possible due to small window. The estimated mean response time can be used the benchmark to meet quality of service (QoS) at end-user. We validate the accuracy of our model using experiments. It is shown that the differences between the results from model and those from experimental are very small below 6 % on average. We also find that the mean response delay for HTTP over SCTP is less than that for HTTP over TCP.

SCTP is a transport layer protocol mainly designed for transferring public switched telephone network signalling over IP networks. Some of the protocols that support data transmission are transmission control protocol (TCP), user datagram protocol (UDP), and stream control transmission protocol (SCTP). SCTP is a newly invented reliable, message-oriented transport protocol. It is similar to TCP and UDP with some additional features such as multi-streaming and multi-homing. Even though it is a new protocol with some advanced features, there is a security loophole issue in data transmission using SCTP. Hence, we proposed a new technique to transmit a secret message by hiding it in a secret matrix using the advanced features of SCTP. The simulation of our proposed method provides better performance and high security for secret message transmission against network attacks.

The Stream Control Transmission Protocol (SCTP) is a transport layer protocol introduced to complement the well known User Datagram Protocol (UDP) and the Transmission Control Protocol (TCP). The current implementation of SCTP heavily leverages some existing mature techniques from TCP. This paper studies the effect that RTOmin (retransmission timeout minimum) has on the selective acknowledgement (SACK), fast retransmission, and timeout mechanisms. We show that there are optimization opportunities in the manipulation of RTOmin while not significantly affecting timeout values and number of fast retransmissions. We demonstrate that RTOmin plays a balancing role in the trigger of timeouts and fast retransmissions and conclude that a dynamically set RTOmin would yield significant time-to-complete (throughput) improvements for SCTP based applications.

The Stream Control Transmission Protocol (SCTP) is a reliable transport protocol to tackle the limitations of TCP and UDP. SCTP was originally designed to transport PSTN signaling messages over IP networks, but is also capable of serving as a general-purpose transport protocol. SCTP provides attractive features such as multi-streaming and multi-homing that may be helpful in high-mobility environment. SCTP congestion control mechanisms are based upon TCP congestion principals with the exception of the fast recovery algorithm. Original SCTP congestion control can perform badly in high-speed wide area networks because of its slow response with large congestion window. We proposed a new congestion control scheme based on the simple congestion window modification for SCTP to improve its performance in high-speed wide area networks. The results of several experiments we performed proved that our new suggested congestion control scheme for SCTP in high-speed network improved the throughput of the original SCTP congestion control scheme significantly.

One of only two new transport protocols introduced in the last 30 years is the Stream Control Transmission Protocol (SCTP). SCTP enables capabilities like additional throughput and fault tolerance for multihomed hosts. An SCTP implementation is included with the Linux kernel and another implementation called sctplib functions successfully in userspace on several platforms but unfortunately neither of these implementations have all of the latest features nor do they perform as well as the FreeBSD kernel implementation of SCTP. We were motivated to produce a portable implementation of the FreeBSD kernel SCTP stack that operates in userspace of any system because of both our desires to obtain a higher performance SCTP stack for Linux as well as to exploit recent developments in hardware virtualization and transport protocol onloading. Unlike any other userspace transport implementation for TCP or SCTP, our userspace SCTP stack simultaneously achieves similar throughput and latency as the Linux kernel TCP stack, without compromising on any of the transport's features as well as maintaining true portability across multiple operating systems and devices. We create a callback API and implement a threshold to control its usage; our userspace SCTP stack with these optimizations obtains higher throughput than the Linux kernel implementation of SCTP. We describe our userspace SCTP stack's design and demonstrate how it gives similar throughput and latency on Linux as the kernel TCP implementation, with the benefits of the new features of SCTP.

The Advanced Simulation and Computing (ASC) Distance Computing (DisCom) Wide Area Network (WAN) is a high performance, long distance network environment that is based on the ubiquitous TCP/IP protocol set. However, the Transmission Control Protocol (TCP) and the algorithms that govern its operation were defined almost two decades ago for a network environment vastly different from the DisCom WAN. In this paper we explore and evaluate possible modifications to TCP that purport to improve TCP performance in environments like the DisCom WAN. We also examine a much newer protocol, SCTP (Stream Control Transmission Protocol) that claims to provide reliable network transport while also implementing multi-streaming, multi-homing capabilities that are appealing in the DisCom high performance network environment. We provide performance comparisons and recommendations for continued development that will lead to network communications protocol implementations capable of supporting the coming ASC Petaflop computing environments.

Currently, the new transport protocol, Stream Control Transmission Protocol (SCTP), is not ECN-capable. The ECN-capable SCTP is proposed in this paper, which is bandwidth-efficient and robust to non-congestion losses. A SCTP source needs to adjust its congestion window when receiving ECN messages. We find the optimal value of the congestion window for a SCTP source in response to ECN messages, and develop a simple and practical method to achieve this optimal congestion window. Our simplified method in achieving the optimal congestion window is attractive because the total goodput performance of SCTP associations or the bottleneck link utilization is not sensitive to the window reduction policies when the network load is heavy. Using complicated methods to fine-tune SCTP or TCP's congestion window in response to congestion indications may be not worth the increase in complexity of the protocol.

We characterize the operation of the SCTP transport protocol over the land mobile satellite channel (LMSC), which is subject to intermittent blockages, and contrast its performance with that of TCP. In particular, we study the ability of SCTP's multihoming capability to adjust to channel blockages. For small files, SCTP's larger connection setup overheads cause single-homed connections to complete transfers slightly slower than TCP. However, depending on the ability of alternate paths, multihoming can improve transfer times so much that SCTP is competitive with an optimized TCP gateway. For large files, quirks in the congestion control of multihomed associations cause performance to deteriorate considerably; though single-homed SCTP is faster than end-to-end TCP, none of the other options can compete with a highly optimized satellite TCP gateway operating without congestion control over the satellite channel. SCTP multihoming semantics do not lend themselves to protocol-spoofing gateways, but some changes to default parameters can help. In the LMSC environment, applications would generally do best to use multihomed SCTP (if available) for small file transfers and gateway-enhanced TCP for larger transactions.

The stream control transmission protocol (SCTP) is a relatively recent transport protocol, offering features beyond TCP. Although SCTP is an alternative transport protocol for the session initiation protocol (SIP), we do not know how SCTP features influence SIP server scalability and performance. To estimate this, we measured the scalability and performance of two servers, an echo server and a simplified SIP server on Linux, for both SCTP and TCP. Our measurements found that using SCTP does not significantly affect data transfer latency. However, the number of sustainable associations drops to 17-21% or to 50% of the TCP value if we adjust the acceptable gap size of packet reordering.

Transport layer reneging

Stream Control Transmission Protocol (SCTP) is an emerging transport protocol for sending data from one point to another over the Internet. The two major new capabilities that are designed into SCTP are the support for multi-homed hosts and the support for multiple streams in a single SCTP association. The purpose of this paper is to investigate the performance of SCTP, measured in terms of association throughput. We analyze the throughput for two experimental scenarios: 1. Transmission of data chunks with different sizes of payload in network stable state (no packet loss). 2. Transmission of data chunks with different sizes of payload in packet loss environment. We have shown the impact of data chunk transmissions with different sizes of payload on the throughput of a multi-homed SCTP association. Our experimental results and analysis show that SCTP provides better throughput in the case of large payload data chunks transmission in both network stable state and packet loss environment.

The stream control transmission protocol (SCTP) is a transport protocol that is robust against network failure. However, it cannot select routes for packet rerouting unless the receiver is a multi-homed host. We propose a method that allows single-homed hosts to support SCTP multi-path access; it allows hosts in a local network to select routes to single-homed severs on the Internet. In this method, a single-homed client constitutes a virtual multi-homed client with a Multi-path gateway (MPGW) and communicates with a single-homed server over multiple routes. Our experiments show that a virtual multi-homed client can communicate with a normal SCTP server and that the method establishes connections that experience the same network failure to recover at different times, whereas IP dynamic routing recovers them at the same time.

The increased number and diversity of underlying networks have made transparent network migration a necessity. Through its support for multi-homing the Stream Control Transmission Protocol (SCTP) enables seamless network mobility by abstracting multiple underlying physical paths into a single end-to-end association. One of these paths is selected as the primary. When a number of retransmission failures occur on the primary path, switchover is initiated to a secondary path. The number of retransmission attempts before switchover is initiated can be configured; however, the delay between each retransmission is managed internally in SCTP using a Retransmission TimeOut (RTO) value. This paper shows that the current SCTP mechanism for calculating RTO values is inappropriate in WLAN environments, since increased Round Trip Times (RTT) significantly distort RTO calculations. Experimental and simulated results indicate that SCTP behaves in a counterintuitive manner which allows more time for switchover as network conditions degrade: delays of up to 187 seconds can be experienced before switchover occurs. We show that additional SCTP parameters need to be carefully configured in order to reduce this switchover delay to a more acceptable level.

The stream control transmission protocol (SCTP) is a reliable transport protocol to tackle the limitations of TCP and UDP. SCTP is specified informally in request for comments (RFC) 2960, but it still lacks a formal specification. Coloured Petri net (CPN) is suitable to modeling a system from the dynamic perspective, and has the ability to analyze it in a formal way. We present a formal model of SCTP association management using the CPN. The model is used to examine some properties of SCTP and to check the internal consistency of RFC 2960. Through CPN Tools analysis, we find a problem with some informal descriptions in RFC 2960 and solve it by modifying our original CPN model.