This special issue aims to direct attention to the great potential offered by virtual shared memory (VSM) on distributed hardware architectures. Industry has not yet exploited the advantages offered by VSMbased tools. This is particularly surprising for two reasons: firstly, much research in this area has been published. Secondly, existing client/serverbased technologies obviously do not fully satisfy all needs of robust distributed applications. Asymmetric application architectures predominate the market. They do not exhibit the advantages offered by distribution: better performance, reliability, and availability through replication of resources and services. In order to overcome these deficiencies, multitier architectures have been introduced to avoid the server bottleneck.
Is it only a matter of time and experience before VSM replaces the messagepassingoriented client/server style which, like the relational model, might succeed as the conceptually superior model, although initially considered too inefficient? Another perspective would be a sensible coexistence of both paradigms, where each enhances the other. VSM could help to overcome the replication deficiencies, whereas distributed object standards like CORBA could offer the necessary vendor interoperability.