110 likes | 184 Views
Explore the design and development of Quality of Service (QoS)-aware containers in Tapas Architecture based on J2EE, with a focus on end-to-end QoS application requirements, core services provisioning, and a case study in a distributed auction scenario.
E N D
TAPAS Architecture: QoS-aware Containers V. Ghini, G. Lodi, N. Mezzetti, F. Panzieri Department of Computer Science University of Bologna
Summary • Goals • An Auction Scenario • Basic QoS services • Tapas Architecture • Future work and references
Goals • Design and development of QoS-aware containers, based on J2EE • This entails: • Deriving end-to-end QoS application requirements from SLAs • Identifying core services for end-to-end QoS provisioning • Case of study: distributed auction scenario
Legend: SLA contract role Auction Scenario TTP CSP Buyer Seller Marketplace / Auctioneer Credit Rating Agency ASP ISP SSP Retail Bank
Auction: non-functional requirements [Adesso02] • Reliability oriented requirements: • High availability • Failure Recovery • Fault tolerance • Performance oriented requirements: • Timeliness of execution • Relative processing speed • Security oriented requirements: • Authentication • Privacy Fairness?
Basic QoS services (1/2) Basic QoS services are necessary to provide end-to-end QoS, independently of the type of application Three principal services: • Configuration Service • Run Time Support Service • Trust Management Service
Basic QoS Services (2/2) • Configuration Service (CS): • Resource Discovery • QoS Negotiation • Resource Reservation • Run Time Support Service (RTSS): • QoS Monitoring • QoS Adaptation • Resource release • termination and advanced garbage collector
Tapas Architecture:Component Execution Environment Application Tier Application Container QoS Contract (SLA) TAPAS Tier 1 2 CS TMS RTSS
Component Execution Environment Application Container + EJB EJB RTSS TMS CS TGCP: -ENS -EPGM J2EE Services TTM Tapas Architecture input output
Future Work and References • Future Work: • design and development of RTSS, CS, and TMS based on input QoS parameters • References • [Adesso02] Warner Beckman, Marina Oleneva “Application Hosting Requirements”, 27 May 2002, Dortmund • [JC02] Jon Crowcroft “Channel Islands in a Reflective Ocean: Large Scale Event Distribution in Heterogeneous Networks”, 23 May 2002, University of Cambridge