Early Lessons from Building Sensor.Network: An Open Data Exchange for the Web of Things

Early Lessons from Building Sensor.Network: An Open Data Exchange for the Web of Things

• 1.
  Early Lessons fromBuilding Sensor.Network: An Open Data Exchange for the Web of Things Vipul Gupta, Poorna Udupi, Arshan Poursohi {vipul.x.gupta, poorna.udupi, arshan.poursohi}@oracle.com First International Workshop on the Web of Things (WoT 2010), PerCom 2010 1
• 2.
  Overview • The Internetis now the Internet of Things ... data storage 2
• 3.
  Overview • The Internetis now the Internet of Things ... data analysis 2
• 4.
  Overview • The Internetis now the Internet of Things ... insights 2
• 5.
  Overview • The Internetis now the Internet of Things ... actions 2
• 6.
  Overview • The Internetis now the Internet of Things ... Sensor.Network: an open data-exchange for the Web of Things 2
• 7.
  Motivation • Interactions withmany SPOTs customers - Sun Modular Data Center, MIT, Warren Wilson College, USGS, Vodafone • Common set of issues: - Data collection - Sharing - Privacy - Search - Collaborative analysis - Notifications • Bring social networking technologies (e.g. tags, annotations, fine-grained access control, web-hooks) to this problem space 3
• 8.
  Motivation • Interactions withmany SPOTs customers - Sun Modular Data Center, MIT, Warren Wilson College, USGS, Vodafone • Common set of issues: - Data collection - Sharing - Privacy - Search - Collaborative analysis - Notifications • Bring social networking technologies (e.g. tags, annotations, fine-grained access control, web-hooks) to this problem space 3
• 9.
  Motivation • Interactions withmany SPOTs customers - Sun Modular Data Center, MIT, Warren Wilson College, USGS, Vodafone • Common set of issues: - Data collection - Sharing - Privacy - Search - Collaborative analysis - Notifications • Bring social networking technologies (e.g. tags, annotations, fine-grained access control, web-hooks) to this problem space 3
• 10.
  Motivation • Interactions withmany SPOTs customers - Sun Modular Data Center, MIT, Warren Wilson College, USGS, Vodafone • Common set of issues: - Data collection - Sharing - Privacy - Search - Collaborative analysis - Notifications • Bring social networking technologies (e.g. tags, annotations, fine-grained access control, web-hooks) to this problem space 3
• 11.
  Benefits • Enables investigationsinto correlations that would otherwise be missed • Enables easy sharing of data and analytical algorithms • Enables data-centric collaboration: editing, annotation, discussion • Frees up domain experts from the chore of setting up and managing an IT infrastructure 4
• 12.
  Sensor.Network Overview • Embraceheterogenous devices • Lower barrier to usage • Easy, RESTful API • Fine-grained access control • General platform for specialized applications 5
• 13.
  “Datastream” Abstraction • A time-series of sensor values sampled together. • Decouples physical sensor from high-level phenomenon of interest. • A datastream may be fed by different sensors at different times • Has associated metadata and privacy controls 6
• 14.
  REST-based API • Sensor.Networkoperations may be accessed via a browser or programmatically using HTTP operations • Most operations require authentication--either password or API Key Datastream creation example: % curl –dump-header hdrs.txt –header “X-SensorNetworkAPIKey: apiKey” –request POST –header “Content- type: application/xml” –data-binary “@Datastream.xml” “http://sensor.network.com/rest/resources/datastreams/” <?xml version=”1.0”?> <datastream> <name>My Pothos Plant</name> <description>Light and temperature readings from my office plant. </description> <tag>light</tag> <tag>temperature</tag> <value><name>light</name> <type>int</type> <units>lumen</units></value> <value><name>temperature</name> <type>float</type> <units>celsius</units> <min>-50.0</min> <max>50.0</max></value> </datastream> 7
• 15.
  REST-based API (moreexamples) • Data insertion % curl –dump-header hdrs.txt –header “X-SensorNetworkAPIKey: apiKey” –request POST –header “Content- type: application/xml” –data-binary “@Data.xml” “http://sensor.network.com/rest/resources/datastreams/id/data” <?xml version=”1.0”?> <sampleData> <sensorNodeId>0014.4F01.0000.01AB</sensorNodeId> <timestamp>2009-07-30T13:31:37.459Z</timestamp> <value>700</value> <value>29.4</value> </sampleData> • Data retrieval % curl –dump-header hdrs.txt –header “X-SensorNetworkAPIKey: apiKey” “http://sensor.network.com/rest/ resources/datastreams/id/data” –header “Accept: text/plain” 8
• 16.
  Security • Authentication • username/password for browser-based access • API key for programmatic access • Authorization uses UNIX-like model based on user classes and access types • HTTPS for confidentiality • Target verification to avoid “alert spamming” 9
• 17.
  “Task specific” Views(Dashboard) • At-a-glance information on datastreams -- health, permissions 10
• 18.
  “Task specific” Views(Map) • Identifies geographical location with additional information in tabbed windows 11
• 19.
  Visualizations • LivePlots: Interactivevisualizations, embeddable in external web pages • Includes standard visualization types: line plots, bar charts, scatter plots • Support for custom visualizations, e.g. dynamic map trace 12
• 20.
  Visualization (cont’d) • Investigating • “Composite” visualizations • Privacy-enhanced visualizations (time-delayed, low-fidelity) 13
• 21.
  Event Notifications • Provides “push” functionality • User can receive email, SMS or HTTP Post (webhook) on events of interest • Events need not be data related, e.g. “alert me when someone adds me to a group” • POST to .../notifications creates a subscription request and returns Id • Target verification 14
• 22.
  Ongoing Work • Data-centriccollaboration: - richer, privacy-aware visualizations, - visual annotation and editing • Integration of statistical packages (e.g. R) as pluggable analysis modules • Expansion of “bi-directional” capabilities: - Device management - General m2m communication, sleep-proxy (IETF CoRE) • Sensor.Network as an application platform: third-party access to user data (delegated authorization) 15
• 23.
  Lessons Learned • Enableinvestigations into correlations that would otherwise be missed • Support device heterogeneity • Open API. Support both PUSH/PULL • Enable easy sharing of data and analytical algorithms • Address privacy concerns • Provide tools for data-centric collaboration, sharing • Lower barrier to usage by domain experts • Hide complex IT infrastructure behind a simple API • Provide simple abstractions 16
• 24.
  Sun SPOT Giveaway •2 Sun SPOT kits (see http://www.sunspotworld.com) available • Email a short (less than 2 pages) proposal to [email protected] with the title “PerCom Giveaway” before noon Mar 31, 2010 • Describe your project • Why Sun SPOTs would be a good match • How you’d share the results (e.g. report, open source, post on http://sunspotworld.com/forums) • What would you do with the kit long term, e.g. rotate among members of a JUG, use to teach a class etc. 17
• 25.
  Q&A 18