Algolia - Hosted Search API
- 1. Instant Search API Build Unique Search Experiences Sylvain Utard VP of Engineering sylvain@algolia.com @sylvainutard Enterprise Search and Analytics
- 2. @algolia Who am I? 5 years @ Exalead, leading the core-engine & NLP teams • C++ • ExaScript (RIP) • Java 2 years @ Algolia, VP of Engineering • C++ • Ruby • Java • and 10+ other languages… @sylvainutard
- 3. @algolia A hosted search API
- 4. @algolia A hosted search API
- 5. @algolia
- 6. @algolia A hosted search API Replies in milliseconds
- 7. @algolia A hosted search API Replies in milliseconds From anywhere
- 8. @algolia A hosted search API Replies in milliseconds From anywhere With intuitive relevance
- 10. @algolia 800+customers in 80+ countries Algolia Today
- 11. @algolia 800+customers in 80+ countries 40B+ Write operationsper month 4B+ User-generated queriesper month Algolia Today
- 12. @algolia Algolia Today 13locations 800+customers in 80+ countries 40B+ Write operationsper month 4B+ User-generated queriesper month
- 14. @algolia Speed matters Half a second delay caused 20% drop in traffic Every 100ms of latency costs them 1% in sales
- 16. @algolia Unique set of constraints High volume of Read & Write operations
- 17. @algolia Unique set of constraints High volume of Read & Write operations High-availability
- 18. @algolia Unique set of constraints High volume of Read & Write operations High-availability Worldwide data distribution
- 19. @algolia API Software Stack Started as a mobile offline SDK Written in C++ Search code embedded in Nginx as a module Indexing is done in a separate process Two redis instances
- 20. @algolia API Hardware Fast CPU (Xeon E5 >3.5GHz) In Memory (128GB) Backed by High-end SSD in Raid-0 (800GB) Specific kernel settings
- 21. @algolia Scaling horizontally Several clusters per location A user is assigned to one master cluster A user can be replicated to N replicate clusters
- 22. @algolia What is a cluster Master-Master Stream of writes via Consensus At least 3 machines
- 23. @algolia A write in practice One of the machines accept the write operation via the API (https) /1/indexes/MyFirstIndex/batch
- 24. @algolia A write in practice The file is saved on the three machines as a temporary file tmp1265 tmp7864 tmp2357
- 25. @algolia A write in practice Launch the consensus by contacting the RAFT master startConsensus(tmp2357, tmp7864, tmp1265)
- 26. @algolia A write in practice 1 -Master send the commit order to all nodes 2- Each node returns the next job ID to master 3- If there is a majority the file is committed
- 27. @algolia A write in practice Same job ID on all hosts Send to slave replicate in parallel Processed in parallel on all hosts job42 job42 job42
- 28. @algolia In case one host is down Continue to accept writes The two other hosts keep jobs Jobs are sequential, will catch up at restart job42job42
- 29. @algolia Distribution Replicate jobs, not the result Send to all machines in parallel Consistent with few seconds delay
- 30. @algolia High availability Multi-regions in one location
- 31. @algolia High availability 13 fully independent locations
- 32. @algolia Network Optimisations API usage moving from servers to browser and mobile apps Get close to end users
- 33. @algolia Distributed Search Network - Worldwide Synchronization
- 34. @algolia Distributed Search Network - Worldwide Synchronization
- 35. @algolia • 13 locations = 25 datacenters • No ideal worldwide provider • AWS is not in India, Eastern EU, Africa… • Need to handle several providers • Anticipate long deliveries / customs • Keep as few providers as possible Distributed Search Network - Worldwide Synchronization
- 36. @algolia DNS is key Used to find the closest location Several DNS providers Good anycast network
- 37. @algolia API Clients DNS health checks are not enough Smart retry logic in all our API Clients
- 38. @algolia Analytics • What are my users searching for? • Top search • Top search without hits • Top refinements • From where do they search for?
- 39. @algolia
- 40. @algolia
- 41. @algolia Analytics • Billions of user-generated queries per month • As-you-type aggregation • ~3 months retentions • Storing all of them in…
- 42. @algolia Analytics • Elasticsearch o/ • … without FTS :) • but with aggregations
- 43. @algolia Analytics• No FTS • No source • Doc values everywhere • SSD only • Custom aggregations (deprecated since ES 1.1.0)
- 44. @algolia Top-k Aggregation • Before • Linear memory consumption • Exhaustivity • After • Constant memory consumption • Approximative but enough
- 45. @algolia Building your worldwide infra - Is long and difficult quest - Is a real asset & differentiator The Future of APIs is Distributed
- 46. @algolia All the details of our architecture are on HighScalability.com Want to know more?
- 47. THANK YOU! sylvain@algolia.com @algolia Build Unique Search Experiences W e are hiring in SF, NYC and Paris 😊