Rate constants for the reactions N( 2 D, 2 P) + C 2 H 2 and C 2 D 2 have been measured using a technique of pulse radiolysis-resonance absorption between 220 and 293 K. Arrhenius parameters have been determined from the temperature dependence of the measured rate constants; the activation energies for the reactions of N( 2 D) were about 0.5 kcal/mol, while those for N( 2 P) were about 0.9 kcal/mol. The H/D isotope effect was found to be very small for both the N( 2 D) + C 2 H 2 and N( 2 P) + C 2 H 2 reactions. The rate constants for N( 2 D) + C 2 H 2 were found to be about 3 times as large as those for N( 2 P) + C 2 H 2 . To understand the overall reaction mechanism of the N( 2 D) + C 2 H 2 reaction, ab initio molecular orbital calculations of the lowest doublet potential energy surface have been performed. It has been found that the initial step of the reaction is the addition of the N atom to the π bond of acetylene. The rate constants have been calculated using conventional transitionstate theory and compared to the experimental results. Possible reaction pathways are discussed on the basis of the ab initio results.

Thermal rate constants for the N(2D,2P) + CH4 (CD4) reactions have been measured using a technique of pulse radiolysis−resonance absorption in the temperature range between 223 and 298 K. Activation energies determined from the temperature dependence were about 1.5 and 1.0 kcal/mol for the reactions of N(2D) and N(2P), respectively. The H/D kinetic isotope effects were about 1.8 and 1.6 for N(2D) and N(2P), respectively. The rate constants for N(2P) + CH4 were much smaller than those for N(2D) by a factor of 40−60. Variational transition-state theory calculations of the rate constants for the N(2D) + CH4 (CD4) insertion reaction have been carried out using the reaction path information obtained from ab initio molecular orbital calculations. The comparison between the calculated and experimental rate constants shows that multiple surface coefficients are larger than the statistical value, meaning that nonadiabatic transitions are important for the N(2D) + CH4 reaction.

The thermal rate constants for the reactions of electronically excited nitrogen atoms, N(2D) and N(2P), with C2H4 and C2D4 have been measured by using a pulse radiolysis−atomic absorption method between 225 and 292 K. From the results of the kinetic isotope effect for the N(2D) reaction, a main reaction mechanism was assigned to be the N addition to the double bond. This conclusion is in accordance with the prediction of our previous ab initio calculations. Variational transition-state theory calculations were performed for the N(2D) reaction by using the results of ab initio molecular orbital calculations. It was suggested that correction of the “multiple surface coefficient” is necessary to interpret the measured rate constants. The ratio of the rate constant of N(2D) to that of N(2P) was found to be close to unity and was much smaller than those for the reactions with H2 and CH4. The deactivation process of N(2P) was determined to be the spin-allowed quenching process, N(2P) + C2H4(S0) → N(4S) + C2H4(T1).

Abstract-If a session initiation protocol (SIP) server receives enough SIP messages to exceed its processing capacity, it goes into overload. When a SIP server receives new SIP messages under overload, the SIP message queue in the server lengthens, which may cause significant throughput degradation in the network. An overload control mechanism using the 503 response code is specified in the SIP standard to improve throughput in the event of such an overload. However, it is known that this is insufficient to control overloading in a large SIP network. In this paper we propose a queue management scheme at a SIP server to improve throughput and prevent overload propagation when overload occurs. In the proposed scheme, SIP messages are dropped based on a dropping probability that is calculated by prioritizing the items in the SIP queue. We also show the effectiveness of the proposed scheme by simulation experiments.Index Terms-IMS, overload control, queue management, SIP.