Tuning SCTP failover for carrier grade telephony signaling

Abstract

The Stream Control Transmission Protocol (SCTP) has not only been selected as the signaling transport protocol of choice in IETF SIGTRAN, the architecture that bridges circuit-switched and IP-based mobile core networks, but also plays a pivotal role in SAE/LTE, the next-generation UMTS/HSPA networks. To meet the redundancy requirements of telecom signaling traffic, SCTP includes a failover mechanism that enables rerouting of traffic from an unreachable network path to a backup path. However, the recommendations provided by IETF on how to configure the SCTP failover mechanism to meet telecom signaling requirements are kept quite general and leave much of the tuning to the telecom equipment vendor and/or operator. Several works by us and others have been carried out to study the effect of different SCTP parameters on the failover performance. The main contribution of this paper is that it gives a coherent treatment of how to configure the SCTP failover mechanism for carrier-grade telephony signaling, and provides practically usable configuration recommendations. The paper also discusses an alternate or complementary way of optimizing the SCTP failover mechanism by relaxing the exponential backoff that foregoes a retransmission timeout in SCTP. Some results showing significantly reduced failover times by use of this mechanism, with only marginal deteriorating effects on a signaling network, are discussed and analyzed in the paper.

Introduction

A large number of technology market research firms forecasts that future telecommunication will, in terms of revenue, be dominated by mobile broadband networks. Today, the mobile broadband market is dominated by 3GPP-based technologies such as UMTS and HSPA. Informa, an independent market analyst firm, forecasts that by year end 2012, there will be 1.3 billion UMTS/HSPA connections representing 78% of the 3G mobile broadband market [24]. In comparison, 56.8 million total connections to mobile WiMAX are anticipated by year end 2012 [24].

A large part of current mobile core networks are built upon legacy TDM- and ATM-based circuit-switched technologies. However, future-generation core networks are expected to be all IP. To make an incremental transition possible, IETF introduced the SIGnaling TRANsport (SIGTRAN) architecture [25]. The SIGTRAN architecture separates voice traffic from control traffic,¹ and specifies how circuit-switched signaling, i.e., Signaling System Nr. 7 (SS7) signaling, is transported over, or terminated in, an IP core network. Currently, around 50% of all shipped SS7 stacks are SIGTRAN stacks, however, this is anticipated to rise up to more than 75% by the end of 2012 [7].

To be able to comply with the requirements of SS7, SIGTRAN introduced a new transport protocol, the Stream Control Transmission Protocol (SCTP). Since its introduction in SIGTRAN, SCTP has also been selected as one of the transport protocols to be used in SAE/LTE, the next-generation UMTS/HSPA network. A key feature of SCTP that sets it apart from TCP and UDP, the two standard Internet protocols, is multihoming. Normally, traffic is sent on a primary network path. However, if this path goes down, traffic is diverted to a secondary, backup, path. To meet the delay and timing requirements of SS7, ITU-T implicitly prescribes that a failover from the primary to a secondary path should take no more than 2 s [16]. However, being considered as a general transport protocol by IETF, SCTP configured as recommended by RFC 4960 [30] exhibits a minimum failover time of 63 s. Several works [14], [20], including IETF [6], have given recommendations on how to configure SCTP for telecom applications, but they are mostly quite vague in their recommendations [6], or focus on one or a few parameters and factors [14], [20]. Many of the findings presented in this paper go back to earlier work by us and others, however, this paper provides a coherent treatment of this subject and translates the results into concrete recommendations. The key factors behind these recommendations are illustrated with experimental results. The paper also reports on a proposal by us that can substantially decrease SCTP failover times by making the backoff factor for the retransmission timer configurable. The possible performance benefit of this proposal is verified by simulation results.

The recommendations in the paper point out that a strict configuration of the SCTP protocol parameters is crucial to reach the signaling application timing requirements. On the other hand, the paper acknowledges the fact that a strict tuning increases the risk of spurious failovers. This is the reason why this tuning has to be carefully conducted. The proposal for a modified RTO policy is motivated as a way to enable a slightly more liberal tuning of the protocol parameters. In the paper, it is shown that by reducing the backoff factor from its default value of 2 to a value in between 1.5 and 2 it is possible to meet the demands of signaling applications using a less strict parameter tuning, and this without violating network stability or performance.

The remainder of the paper is organized as follows: Section 2 gives an overview of SCTP, focusing on multihoming and the failover mechanism. Some of the results presented in this paper are illustrated with experiments, and Section 3 describes the common parts of the experiment setups used in these experiments. The next two sections constitute the core of the paper; Section 4 discusses how to tune SCTP protocol parameters to make SCTP meet telecom signaling failover requirements, while Section 5 considers the impact of external factors, such as network and traffic characteristics, on the SCTP failover performance. Next, Section 6 surveys our proposal for a new backoff mechanism. The paper concludes in Section 7 by summarizing the configuration recommendations given in the paper and recommending avenues for future work.