The Fifth Elephant 2017

On data engineering and application of ML in diverse domains

Search Infrastructure @ Slack using Lambda Architecture

Submitted by Ananth Durai (@vananth22) on Thursday, 27 April 2017

videocam
Preview video

Technical level

Intermediate

Section

Full talk for data engineering track

Status

Submitted

Vote on this proposal

Login to vote

Total votes:  +9

Abstract

Slack is a collaboration tool for teams. We’re on a mission to make your working life simpler, more pleasant, and more productive. Search is the core feature of Slack offerings as Slack itself is an acronym for “Searchable Log of all conversation & knowledge”.
At Slack, we experiment frequently with various machine learning models to improve search experience so rebuilding search indexes are critical for search infrastructure.
This talk will centre around lambda architecture, common pitfalls of lambda architecture and best practices, an overview of Slack search infrastructure and our experience in building Solr offline indexing at scale.

Outline

This talk will be around,
1. Lambda Architecture
2. Common pitfalls of Lambda Architecture
3. Design patterns to handle common pitfalls of Lambda Architecture
4. Apache Solr offline indexing
5. Apache Solr search infrastructure.

Requirements

Some basic understanding of mapreduce and stream processing.

Speaker bio

Ananth Packkildurai is currently working as a Senior Data Engineer at Slack Technologies, San Francisco and has over thirteen years of experience in building systems at scale. Prior to Slack, Ananth worked as a senior data engineer at Bazaarvoice Inc to build large scale consumer reviews analytical platform. He works closely with Hadoop, Apache Crunch, Kafka, Apache Solr, Apache Spark, Druid and other big data platform tools.
Linkedin: https://www.linkedin.com/in/ananthdurai

Links

Slides

https://speakerdeck.com/vananth22/search-infrastructure-using-lambda-architecture

Preview video

https://youtu.be/pzikbzGSqgM

Comments

  • 1
    Zainab Bawa (@zainabbawa) Reviewer a year ago

    Thanks for submitting this proposal, Ananth. Have you presented this talk earlier? We require draft slide deck detailing the content that you will cover in the talk and structure of the presentation. You also need to upload a two-min preview video explaining what this talk is about, and what is the key takeaway for participants.

    • 1
      Ananth Durai (@vananth22) Proposer a year ago

      Hi Zeinab, I’ve uploaded the draft slide and a link from Slack engineering blog that talks about search in general. This talk will touch upon some of the aspect mentioned in the blog but mostly centered around data infrastructure. I will follow up with short 2 min video on it.

      • 1
        Zainab Bawa (@zainabbawa) Reviewer a year ago

        Ananth, update on the preview video? We are closing decisions on proposals on 10 June. The editorial team needs the video before 10 June.

        • 1
          Ananth Durai (@vananth22) Proposer a year ago

          Hi there,
          I’ve updated a preview video and requirement for the talk as well.

  • 1
    Damon Young (@damonyoungdf) a year ago

    Hi, Ananth. Thanks for the great talk. Can I ask: why did you choose Solr over, say, Elastisearch, for the main search platform at Slack?

Login with Twitter or Google to leave a comment