Upload
coldbeans-software
View
436
Download
0
Embed Size (px)
Citation preview
On data model for context–aware services
Dmitry NamiotLomonosov Moscow State University
Network proximity
• A special model for context-aware services
• Context described as a set of wireless networks (nodes)
• Wi-Fi access points, Bluetooth nodes, Bluetooth tags
• Data could be directly associated with network nodes.
Context
• Context is anything we can add to location• Models for context-aware systems:• Data exchange depending on the context• Situational awareness• Context-aware data discovery and data
search
Network nodes
• Wi-Fi (Bluetooth) devices are everywhere• Wi-Fi (Bluetooth) is presented in every
mobile phone• Easy to measure (existing standards)• We can reuse existing infrastructure• There is no connection with location (geo-
coordinates)
Metrics
• The basic element: fingerprint• A list of “visible” nodes: ID, MAC-address,
RSSI (signal strength)• Occurrence counting• RSSI-based “distance”
iBeacons
Related projects
Related projects
CityProximus: prototype
• The prototype (Fraunhofer, FOCUS):
CityProximus:shortly
• A set of rules:• Network fingerprint – > data chunks
(information)• Can use existing networks as well as
especially created wireless nodes
CityProximus: architecture
• Data base for network proximity rules and content
• Rules editor• Application server (API for developers)• Mobile application for access to content
(context-aware browser)
CityProximus: data modelRules: productionsIf (fingerprint condition) then { present some
content }RETE algorithmREST API with JSON output:[ { “type”:”some_type”,”data”:”some_data”}, {“type”: ...},...]
CityProximus: NoSQL model
• Key-value store• Key: MAC-address• Value: JSON array• Apache Accumulo• Query by index• Bloom filter for cache