Improving the Scalability of Fault-Tolerant
Database Clusters: Early Results
M. Patiño-Martínez, B. Kemme, G. Alonso
Replication has become a central element in
modern information systems playing a dual role: increase availability
and enhance scalability. Unfortunately, most existing protocols
increase availability at the cost of scalability. This paper presents
architecture, implementation and performance of a middleware based
replication tool that provides both availability and better
scalability than existing systems. Main characteristics are the usage
of specialized broadcast primitives and efficient data propagation.
Proc. of the IEEE 22nd Int. Conf. on Distributed Computing Systems 2002,
ICDCS'02. Vienna, Austria. July 2002.
Click to get the PostScript , Gzipped PostScript.