Download presentation
Presentation is loading. Please wait.
1
Searching for Rio: Azure Search, NBC Sports, and the Olympics
Microsoft 2016 11/15/ :31 PM BRK2183 Searching for Rio: Azure Search, NBC Sports, and the Olympics Christopher Stone Software Engineer © 2016 Microsoft Corporation. All rights reserved. MICROSOFT MAKES NO WARRANTIES, EXPRESS, IMPLIED OR STATUTORY, AS TO THE INFORMATION IN THIS PRESENTATION.
2
Agenda Azure Search overview Customer case study Code walkthrough
11/15/ :31 PM Agenda Azure Search overview Customer case study Code walkthrough © 2014 Microsoft Corporation. All rights reserved. Microsoft, Windows, and other product names are or may be registered trademarks and/or trademarks in the U.S. and/or other countries. The information herein is for informational purposes only and represents the current view of Microsoft Corporation as of the date of this presentation. Because Microsoft must respond to changing market conditions, it should not be interpreted to be a commitment on the part of Microsoft, and Microsoft cannot guarantee the accuracy of any information provided after the date of this presentation. MICROSOFT MAKES NO WARRANTIES, EXPRESS, IMPLIED OR STATUTORY, AS TO THE INFORMATION IN THIS PRESENTATION.
3
Azure Search overview 11/15/2018 12:31 PM
© 2014 Microsoft Corporation. All rights reserved. MICROSOFT MAKES NO WARRANTIES, EXPRESS, IMPLIED OR STATUTORY, AS TO THE INFORMATION IN THIS PRESENTATION.
4
Typical search scenarios
11/15/ :31 PM Typical search scenarios eCommerce Product search Newsfeed Newer articles typically more relevant Geospatial search Location-aware data Unstructured corpora Index large PDFs and MS Office files © 2014 Microsoft Corporation. All rights reserved. MICROSOFT MAKES NO WARRANTIES, EXPRESS, IMPLIED OR STATUTORY, AS TO THE INFORMATION IN THIS PRESENTATION.
5
Why Azure Search? Scale Relevancy Natural language processing
11/15/ :31 PM Why Azure Search? Scale Easily add more or fewer resources from the Azure Portal or management API Relevancy Connect search results to business goals with great control over search ranking Natural language processing Leverage Microsoft’s deep knowledge of natural language processing Integration Load and update automatically with integrated indexers Analytics View search insights from PowerBI © 2014 Microsoft Corporation. All rights reserved. MICROSOFT MAKES NO WARRANTIES, EXPRESS, IMPLIED OR STATUTORY, AS TO THE INFORMATION IN THIS PRESENTATION.
6
Core search functionality
11/15/ :31 PM Core search functionality Index schema The shape of your data Data types String, StringCollection, Int32, Int64, Double, Boolean, DateTime, GeoPoint Indexing features Searchable, Filterable, Sortable, Retrievable © 2014 Microsoft Corporation. All rights reserved. MICROSOFT MAKES NO WARRANTIES, EXPRESS, IMPLIED OR STATUTORY, AS TO THE INFORMATION IN THIS PRESENTATION.
7
Core search functionality
11/15/ :31 PM Core search functionality Analyzers Built in language support for 50+ languages Customize of word breaking and other text pre-processing tasks Suggesters Auto-complete common search queries E.g. product titles Scoring Affect result ranking beyond just textual similarity Field weighting Boost newer content Boost content closer to a given geographic point © 2014 Microsoft Corporation. All rights reserved. MICROSOFT MAKES NO WARRANTIES, EXPRESS, IMPLIED OR STATUTORY, AS TO THE INFORMATION IN THIS PRESENTATION.
8
Loading your content Indexers Content sources Content types Push API
11/15/ :31 PM Loading your content Indexers Automatically connect to your content in the cloud Load and update your search index on a schedule Content sources Blob, Table, DocumentDB, Azure SQL Content types SQL, SProc, JSON, CSV, HTML, PDF, MS Office Push API Index content as soon as it is available REST service Batch processing © 2014 Microsoft Corporation. All rights reserved. MICROSOFT MAKES NO WARRANTIES, EXPRESS, IMPLIED OR STATUTORY, AS TO THE INFORMATION IN THIS PRESENTATION.
9
Finding your content Filters Facets Highlighting More like this
11/15/ :31 PM Finding your content Filters Limit your query to documents matching specific metadata properties Facets Aggregate sum of documents in a particular category Drilldown into more specific results Highlighting Preview each result document with search terms highlighted More like this Return results similar to a given document © 2014 Microsoft Corporation. All rights reserved. MICROSOFT MAKES NO WARRANTIES, EXPRESS, IMPLIED OR STATUTORY, AS TO THE INFORMATION IN THIS PRESENTATION.
10
Lucene query language Term boosting Regular expressions Fuzzy search
11/15/ :31 PM Lucene query language Term boosting Give weight to specific search terms search=quiet^2 clean^3 +hotels Regular expressions Enables infix and suffix queries search=/[hm]otel/ Fuzzy search Based on Levenshtein edit distance algorithm search=atlanta~1 +hotels © 2014 Microsoft Corporation. All rights reserved. MICROSOFT MAKES NO WARRANTIES, EXPRESS, IMPLIED OR STATUTORY, AS TO THE INFORMATION IN THIS PRESENTATION.
11
Scalability Partitions Replicas Search units
11/15/ :31 PM Scalability Partitions Each partition holds up to 15M or 60M documents, depending on SKU Add more partitions as your corpus grows Up to 1.4 billion documents total on the S3 SKU (currently in preview) Replicas Each replica serves up to approximately 15, 60, or 120 queries per second, depending on SKU Add more replicas as your search traffic increases Search units The unit of billing Equal to num_partitions x num_replicas © 2014 Microsoft Corporation. All rights reserved. MICROSOFT MAKES NO WARRANTIES, EXPRESS, IMPLIED OR STATUTORY, AS TO THE INFORMATION IN THIS PRESENTATION.
12
Demo Search samples
13
Case Study: NBC Sports 11/15/2018 12:31 PM
© 2014 Microsoft Corporation. All rights reserved. MICROSOFT MAKES NO WARRANTIES, EXPRESS, IMPLIED OR STATUTORY, AS TO THE INFORMATION IN THIS PRESENTATION.
14
Requirements Content originates on premises Video results
11/15/ :31 PM Requirements Content originates on premises Expose articles and athlete data to search Video results Dedicated results for streaming and replay videos Synonyms and misspellings Athlete names have many variants, when entered by end users Entity matching Provide “top hit” results for specific athletes, sports, teams Type-ahead suggestions Suggest queries and result pages for partial text matching athlete names, venues, sports © 2014 Microsoft Corporation. All rights reserved. MICROSOFT MAKES NO WARRANTIES, EXPRESS, IMPLIED OR STATUTORY, AS TO THE INFORMATION IN THIS PRESENTATION.
15
Demo Frontend experience
16
11/15/ :31 PM Architecture © 2014 Microsoft Corporation. All rights reserved. MICROSOFT MAKES NO WARRANTIES, EXPRESS, IMPLIED OR STATUTORY, AS TO THE INFORMATION IN THIS PRESENTATION.
17
Scoring Freshness Weighting
11/15/ :31 PM Scoring Freshness Quadratic boost to boost to content published within the last 2 days Boosting effect decreases slowly from the first result; quickly from later results Weighting Apply additional boost to matches in Title and Tag fields More specific data typically has more information value © 2014 Microsoft Corporation. All rights reserved. MICROSOFT MAKES NO WARRANTIES, EXPRESS, IMPLIED OR STATUTORY, AS TO THE INFORMATION IN THIS PRESENTATION.
18
Demo Scoring before & after
19
Query volume Total queries Avg queries per day Avg search latency
11/15/ :31 PM Query volume Total queries Avg queries per day Avg search latency © 2014 Microsoft Corporation. All rights reserved. MICROSOFT MAKES NO WARRANTIES, EXPRESS, IMPLIED OR STATUTORY, AS TO THE INFORMATION IN THIS PRESENTATION.
20
Demo Query analytics dashboard
21
Text analytics Content clustering Top queries
11/15/ :31 PM Text analytics Content clustering Identify common topics among source content articles A topic is a general grouping of commonly co-occurring words Top queries Identify topic terms most commonly associated with the most frequent queries Uncover semantic similarity between queries and documents Top queries with no results Identify common misspellings and add them to the synonyms list Identify gaps in content availability © 2014 Microsoft Corporation. All rights reserved. MICROSOFT MAKES NO WARRANTIES, EXPRESS, IMPLIED OR STATUTORY, AS TO THE INFORMATION IN THIS PRESENTATION.
22
Walkthrough
23
Architecture Content lives on blob storage
Indexer loads the content every 5 minutes
24
Demo Building our search engine Microsoft 2016 11/15/2018 12:31 PM
© 2016 Microsoft Corporation. All rights reserved. MICROSOFT MAKES NO WARRANTIES, EXPRESS, IMPLIED OR STATUTORY, AS TO THE INFORMATION IN THIS PRESENTATION.
25
Free IT Pro resources To advance your career in cloud technology
Microsoft Ignite 2016 11/15/ :31 PM Free IT Pro resources To advance your career in cloud technology Plan your career path Microsoft IT Pro Career Center Cloud role mapping Expert advice on skills needed Self-paced curriculum by cloud role $300 Azure credits and extended trials Pluralsight 3 month subscription (10 courses) Phone support incident Weekly short videos and insights from Microsoft’s leaders and engineers Connect with community of peers and Microsoft experts Get started with Azure Microsoft IT Pro Cloud Essentials Demos and how-to videos Microsoft Mechanics Connect with peers and experts Microsoft Tech Community © 2016 Microsoft Corporation. All rights reserved. MICROSOFT MAKES NO WARRANTIES, EXPRESS, IMPLIED OR STATUTORY, AS TO THE INFORMATION IN THIS PRESENTATION.
26
Please evaluate this session
11/15/ :31 PM Please evaluate this session Your feedback is important to us! From your PC or Tablet visit MyIgnite at From your phone download and use the Ignite Mobile App by scanning the QR code above or visiting © 2014 Microsoft Corporation. All rights reserved. MICROSOFT MAKES NO WARRANTIES, EXPRESS, IMPLIED OR STATUTORY, AS TO THE INFORMATION IN THIS PRESENTATION.
27
11/15/ :31 PM © 2014 Microsoft Corporation. All rights reserved. MICROSOFT MAKES NO WARRANTIES, EXPRESS, IMPLIED OR STATUTORY, AS TO THE INFORMATION IN THIS PRESENTATION.
Similar presentations
