Analysis of Performance and Fault-Tolerance of Database Management Schemes for PCS
Dipak Ghosal · 1995
This paper presents the performance and fault-tolerance analysis of four different placement schemes for the HLR (Home Location Register) and the VLR (Visitor Location Register) databases in the signalling network. Based on a mean value analysis it has been shown that distributing the HLR and the VLR database may be required for high degree of penetration of PCS usage. A simple distributed scheme where the HLR is maintained in the Service Control Point (SCP) and VLR is co-located (physically or logically) in the SP (Signalling Point) can significantly reduce the impact of PCS signalling on the existing signalling network. Performance of distributed architectures with replication depends largely on how the databases are replicated. The results show that performance gains are achievable when the replication is based on the roaming characteristics of the mobile station. Fault-tolerance of the different placement strategies is based on a simple failure and recovery model of the databases. The results show that the system can be made more robust against database failures with a distributed and replicated database placement strategy if all the distributed replicated copies are always consistent with each other. This may incur additional performance penalties. The analysis is based on a number of simplifying assumptions and is only intended for a comparative study of different database placement schemes.