Reducing session establishment delay using timed out packets in SIP signaling network

Abstract

SUMMARYSession Initiation Protocol (SIP) has a retransmission mechanism to maintain the reliability for its real time transmission. But these real time transmissions cause overload in the server and creates redundant messages. SIP does not offer sufficient mechanisms for handling overload situations. In this paper, we study the SIP system behavior by separating signaling traffic in two different classes (I nvites and Non-invites) and creating a cut-off priority queueing model. SIP retransmission mechanism with timeout is modeled as a queueing system with impatient customers. Using this model, the effect of unnecessary retransmissions is studied, and delay distribution and loss probability (P b_loss) are calculated. The proposed analytical model is verified with simulations that demonstrate that the inclusion of timeout gives better delay performance. Using P b_loss, an algorithm is developed to control the overload in hop-by-hop transaction as described in RFC 6357. The simulation results for the proposed model with overload control and the standard SIP as per RFC 3261 are compared. The results demonstrate that the server utilization factor of an overloaded server always remains less than one and hence avoids system collapse. Further, the redundant messages in the system are reduced by 30% as compared with the standard SIP network.

On fluid‐flow modeling of priority based request scheduling for finite buffer SIP server

“…The presented distributed overload‐control mechanism uses transaction response time as the basis for implicit overload detection. Mishra et al studies the SIP system behavior by separating signaling traffic in 2 different classes (INVITE and non‐INVITE) and uses a priority queuing model. To reduce the rate of retransmissions, requests exceeding a certain timeout are removed from the queue, when an upstream server receives a feedback information indicating the overload situation from the downstream server.…”

“…In the literature, SIP messages are prioritized according to their roles during the SIP session establishment, such as higher priority is given to INVITE messages over non‐INVITE messages and vice versa . In Guduru and Usha, the overload control is achieved with a priority queuing model to reduce the retransmissions by giving high priority to retransmitted requests and to avoid transmitting unnecessary retransmissions towards downstream of SIP network.…”

“…In the literature, SIP messages are prioritized according to their roles during the SIP session establishment, such as higher priority is given to INVITE messages over non-INVITE messages and vice versa. 20,[32][33][34] In Guduru and Usha, 35 the overload control is achieved with a priority queuing model to reduce the retransmissions by giving high priority to retransmitted requests and to avoid transmitting unnecessary retransmissions towards downstream of SIP network. However, the retransmission identification approach (based on call-Id) used in Guduru and Usha 35 does not comply with the SIP standards, since call-Id does not identify a transaction but a session.…”

A load scheduler for SIP proxy servers: design, implementation and evaluation of a history weighted window approach

“…Because of SIP's popularity and rapidly growing deployments in the Internet Protocol (IP) networks, the SIP server overload [5][6][7][8][9][10][11][12], its security [1,2,[13][14][15][16] and session mobility [17][18][19] are the main challenges of the network. Mishra et al presents a cut-off priority queuing scheduling with a timeout method [5]. In [8], two PI controllers are used to propose an overload control system, for adjusting the queue length and the proxy servers' CPU load.…”

“…r 1 is the second WART which is calculated using (4). This will be repeated in (5). To show the recursive relation, the calculation of r 3 is shown in the following with more details (6).…”