Upgrade to Pro — share decks privately, control downloads, hide ads and more …

Instant browsing for static sites with Algolia

Josh Dzielak
December 05, 2017

Instant browsing for static sites with Algolia

The "R" in Google's RAIL web performance model is "Response". A response to a user's action must take place in less than 100ms for the experience to feel immediate and natural. When content must be fetched from a traditional backend or API in response to a user's interaction, it can be very difficult to meet that requirement. Algolia can help. Designed to power low-latency search-as-you-type experiences, Algolia responds to content fetching requests as "fast as ping". Learn how Algolia works and see how you can speed up your static sites. Demo included.

Josh Dzielak

December 05, 2017
Tweet

More Decks by Josh Dzielak

Other Decks in Technology

Transcript

  1. Josh Dzielak
    Instant browsing for
    static sites with Algolia
    Developer Relations Lead, Algolia
    12/05/2017
    Serverless London Meetup
    @dzello
    github.com/dzello

    View full-size slide

  2. Building search, the Algolia way
    @dzello · @algolia · @ServerlessLDN

    View full-size slide

  3. The multiple faces of search
    Instant search Autocomplete
    Faceting
    Full text search
    @dzello · @algolia · @ServerlessLDN

    View full-size slide

  4. “...search without a search bar?!”

    View full-size slide

  5. Search doesn’t always need keywords
    > Browsing
    > Discovery
    > Navigation
    @dzello · @algolia · @ServerlessLDN

    View full-size slide

  6. Speed is crucial to browsing and
    navigation experiences that users enjoy
    @dzello · @algolia · @ServerlessLDN

    View full-size slide

  7. Avoid the LSOD

    View full-size slide

  8. Source: https://blog.algolia.com/performant-web-animations/
    “Developers often invest quite a bit of time to reduce first
    page loads by even a few milliseconds, but forget to
    consider the impact of the interactions that follow.”
    — Emily Hayman, Solutions Engineer, Algolia
    @dzello · @algolia · @ServerlessLDN

    View full-size slide

  9. Source: https://developers.google.com/web/fundamentals/performance/rail
    “The majority of time users spend in your site isn't waiting
    for it to load, but waiting for it to respond while using it.”
    — Google’s RAIL web performance model
    Especially true for static sites!
    @dzello · @algolia · @ServerlessLDN

    View full-size slide

  10. GOOGLE RAIL MODEL

    View full-size slide

  11. Algolia = fast like ping
    @dzello
    @algolia
    Data for millions of searches on a large dataset
    designed for low-latency search-as-you-type experiences

    View full-size slide

  12. C++, NGINX & bare metal
    @dzello
    @algolia
    Searches start and end inside of an NGINX module

    View full-size slide

  13. Reduce network latency (like CDN)
    @dzello · @algolia · @ServerlessLDN

    View full-size slide

  14. ⚡ Demo ⚡
    Algolia vs. “Backend”
    https://algolia-vs-backend.glitch.me/

    View full-size slide

  15. Bonus: FaaS indexing w/ cloud functions
    @dzello · @algolia · @ServerlessLDN

    View full-size slide

  16. Josh Dzielak
    Thank you!
    Developer Relations Lead, Algolia
    12/05/2017
    @dzello
    github.com/dzello
    Serverless London Meetup

    View full-size slide