Author_Institution :
Dept. of Comput. Sci., Soongsil Univ., Seoul, South Korea
Abstract :
In service-oriented computing, service consumers invoke services deployed by service providers, and service providers are liable to manage the quality of their services including scalability, availability and performance. Especially, scalability is hard to achieve, mainly due to the unknown nature of service consumers and unpredictable volumes of service invocations. Services with poor scalability would lose the consumer bases, resulting in low revenues. Conventional approaches to scalability of network, database, and distributed computing are hardware-oriented solutions. In this paper, we present software-oriented approaches to assuring high scalability of services in cloud computing. We first define scalability in service-oriented computing, present formal foundation of services scalability. Then, we present two effective scalability assuring schemes; service replication and migration. The proposed schemes are distinguished from conventional scalability schemes in three ways; tailored for service-oriented computing, software solutions rather than hardware solutions, and being implementable with current SOA and languages. To show the applicability of proposed schemes, we show a result of implementing the schemes and experiments with them. Using the methods, services scalability can be greatly enhanced dynamically by using software solutions.
Keywords :
cloud computing; service-oriented architecture; SOA; cloud computing; high scalability assurance; service migration; service replication; service-oriented computing; software-oriented approach; Degradation; GSM; Measurement; Quality of service; Scalability; Software; Time factors; Cloud Computing; Migration; Scalability;