An important issue in the design of future Personal Communication Services
(PCS) networks is the efficient management of location information. The cur
rent IS-41 standard PCS architecture uses a centralized database, the Home
Location Register (HLR), to store service and location information of each
mobile registered in the PCS network. If the HLR fails, all incoming calls
to a mobile from hosts which are not in the same location area as the mobil
e are lost. Location updates from mobiles to the HLR are also lost. Once th
e HLR is functional it cannot direct calls to mobiles immediately as mobile
s could have changed their location during the HLR's failure. Fast recovery
from a failure of the HLR is hence important. A link failure in the networ
k could partition the network resulting in a loss of location updates from
mobiles affected by the failed link. We present a new protocol for fast rec
overy of the HLR after a HLR failure or an intermediate link failure. The p
rotocol does not require use of wireless bandwidth during the recovery proc
ess, has a bounded recovery period and is simple to implement making it an
appealing choice in the design of future mobile networks. We analyze the pr
otocol in order to find a medium between protocol cost and the recovery int
erval. We use the protocol to recover from VLR failures. We then use the pr
otocol to recover from database failures in a distributed database architec
ture. (C) 2000 Elsevier Science B.V. All rights reserved.