Social Monitoring Tool codename Looking Glass, Patrice Pelland

913 views

Published on

Published in: Technology
0 Comments
1 Like
Statistics
Notes
  • Be the first to comment

No Downloads
Views
Total views
913
On SlideShare
0
From Embeds
0
Number of Embeds
1
Actions
Shares
0
Downloads
8
Comments
0
Likes
1
Embeds 0
No embeds

No notes for slide

Social Monitoring Tool codename Looking Glass, Patrice Pelland

  1. 1. Incubation: Looking glass - Progression to Windows Azure Patrice Pelland Principal Dev Manager, Microsoft Corp.
  2. 2. Social Monitoring – Incubation • Social Monitoring incubation project codename Looking Glass • Silverlight 4 application, Windows Phone 7 application and iPhone/iPad application – Code Reuse. • Progression of the back end from • Windows Server 2008 R2 with WCF Services, SQL 2008 R2 • Windows Azure (Web Roles and Worker Roles), SQL Azure • Windows Azure (Web Roles and Worker Roles), SQL Azure, Azure Storage (Table Store, Blob storage)
  3. 3. WP 7 Application
  4. 4. Code Re-Use Explorations • Since Silverlight is X-Platform and X-Browsers we were able to have our project run (with the SL download) across browsers and cross OS. • Worked on PCs, Mac OSx and Linux (via Moonlight) • Worked on iPhone/iPad via Monotouch • Kept 85% of the code to have the code run on iPhone/iPad and Linux.
  5. 5. Back End Progression – Phase 1
  6. 6. Why it did not scale? • Regular Social APIs with limitations on # of calls • Generic Data Acquisition Service • No partitioning
  7. 7. Back-End Progression – Phase 2
  8. 8. Scalability issues still there for the back end… • Crawling became more scalable with worker queues • Still not the complete data corpus and not fast enough to query – Full text index and tag indexes
  9. 9. Bing Data Agg.LG Data Aggregators LG SQL Azure Bing Indexers LG Indexers Sentiment Processing CDN (xaps, jpg, png, css, …)Web Roles WCFWeb Roles LG Worker Role Media Posting Incubation v2.0 Phase 3 Visualization Services LG – LMI Data Agg. Bing + LG Data AggregatorsInternet LG Azure Bing Azure Sentiment Processing Azure Blob Storage – Backup Index
  10. 10. Conclusion… • Moving to Azure was cost effective • Reliable, scalable and available • Changing the approach to using indexers gave us a more interesting and complex query and filtering capability.

×