A reliable — double edged strategy for HLR mobility database failure detection and recovery in PCS networks

Since the main objective of PCS networks is to provide ldquoanytime-anywhererdquo cellular service, the existence of lost calls has become the major problem that hardly degrades the performance of such networks. A lost call arises when the location information of mobile users stored in Home Location Register (HLR) mobility database is corrupted or outdated. In order to tackle this situation, it is becoming necessary to tolerate HLR failure by implementing efficient recovery schemes. Traditionally, HLR mobility database is periodically checkpointed. After failure, HLR database is restored by reloading the backup. However, checkpointing suffers from major two hurdles, which are; (i) choosing the optimal checkpointing interval is still a challenge, and (ii) backup information may be obsolete. This paper proposes a reliable HLR failure detection and recovery strategy built upon a double edged architecture of HLR mobility database. The main contribution is to introduce novel registration and call delivery schemes for PCS networks. The proposed strategy is fault tolerant as it examines the availability of both the main HLR of the system as well the backup (BC) continuously, then take the suitable actions to keep the system working with relatively no lost calls. On the other hand, our architecture is double edged as both HLR and BC are employed in the system operation all the time. Experimental results have shown that the proposed failure detection and recovery strategy outperforms traditional checkpointing techniques as it introduces better performance with relatively no lost calls.