Context caches in the Cloud
-
Upload
said-oezcan -
Category
Engineering
-
view
311 -
download
3
description
Transcript of Context caches in the Cloud
Context caches in the Clouds
October 2, 2014M. Said Özcan
by Saad Liaquat Kiani, Ashiq Anjum, Nick Antonopoulos, Kamran Munir and Richard McClatchey
Table of Contents
• Introduction
• The context provisioning architecture
• Context cache
• The bipartite context cache
• Cloud based evolution
• Conclusion
Introduction
• Context (information related to user. i.e user's personal situations)
• Temporal aspect (e.g. valid for a certain duration)
• Different validity durations and access frequencies
• Investigating various context-caching strategies and proposing novel bipartite caching mechanism
• Anything, anytime and anywhere (ubiquitous systems)
• Large number of users, devices, data sources and services
The context provisioning architecture
The context provisioning architecture (cont.)
CxC (Consumer) • Uses context data • Retrieve information by, (a) send subscription to CxB (b) direct on-demand query
CxB (Broker) • Main coordinating component of the architecture • Cache produced context
CxP (Provider) • Provides contextual information • Gathers data from a collection of sensors, network/cloud services • Usually specialized in a particular context domain (e.g. location)
The context provisioning architecture (cont.)• Improving the scalability with federation of multiple context brokers • Providing location transparency to the local clients (CxCs and CxPs) of each broke • Routing of context queries/subscriptions and responses/notifications across
distributed brokers, discovery and lookup functions • Context information is represented in ContextML (XML based schema)
The context provisioning architecture (cont.)
• ContextML, context data relates to an entity (a user, a username, a SIP or an email address etc..) and is of a certain scope (e.g. weather, location, activity and user preferences)
• Also a temporal validity is associated with ContextML encoded context data through the timestamp and expiry tags
• A parser, ContextML Parser, has been implemented as a Java library for Java SE, EE and the Android platforms
• Conceptual diagram of how the system components may operate in a federation of context brokers shown in Figure 3
The context provisioning architecture
Context cache
Context cache (cont.)
Context cache (cont.)
Context cache (cont.)
The bipartite context cache
Cloud based evolution
Conclusion
"The Context Provisioning Architecture employs a caching mechanism at the context broker, which positively affects the mean query satisfaction time between context consumers and
providers."
Thanks