Learning Library

← Back to Library

Asynchronous Cloud Services for Scalable Ads

Key Points

  • Leveraging cloud infrastructure gives the platform unmatched stability and near‑perfect uptime for customers.
  • All services that can be made asynchronous are placed on an event‑stream queue, decoupling heavy backend processing from the front‑end and preserving responsive user interactions.
  • A standardized micro‑service development and deployment pipeline, together with integrated monitoring and metrics, makes the system secure, plug‑and‑play, and easy to manage.
  • The service mesh provides a uniform way to route and block traffic between services, logging erroneous events to aid in tracing and mitigating attacks such as DDoS.
  • By distributing the architecture globally in the cloud, the platform can serve ads in roughly 100 ms, meeting strict latency goals.

Full Transcript

# Asynchronous Cloud Services for Scalable Ads **Source:** [https://www.youtube.com/watch?v=utZU8x2wfoY](https://www.youtube.com/watch?v=utZU8x2wfoY) **Duration:** 00:01:52 ## Summary - Leveraging cloud infrastructure gives the platform unmatched stability and near‑perfect uptime for customers. - All services that can be made asynchronous are placed on an event‑stream queue, decoupling heavy backend processing from the front‑end and preserving responsive user interactions. - A standardized micro‑service development and deployment pipeline, together with integrated monitoring and metrics, makes the system secure, plug‑and‑play, and easy to manage. - The service mesh provides a uniform way to route and block traffic between services, logging erroneous events to aid in tracing and mitigating attacks such as DDoS. - By distributing the architecture globally in the cloud, the platform can serve ads in roughly 100 ms, meeting strict latency goals. ## Sections - [00:00:00](https://www.youtube.com/watch?v=utZU8x2wfoY&t=0s) **Asynchronous Cloud Service Mesh** - The speaker explains how using an asynchronous, event‑stream‑based micro‑service architecture together with a secure service mesh provides high uptime, decoupled backend processing, easy monitoring, and sub‑100 ms ad delivery. ## Full Transcript
0:00[Music] 0:01the cloud gives us great stability and 0:04what that provides to our customers is 0:06uptime that's unparalleled one of the 0:10key aspects of how we design our market 0:12services that has allowed us to be as 0:14flexible and successful as we have is 0:17that any service that we can make 0:20asynchronous we make asynchronous and we 0:23put everything in event stream queue and 0:26that allows us to decouple the backend 0:29services that do the heavy lifting from 0:31the front-end services that provide 0:33customer interaction so that way their 0:36customer interaction isn't degraded 0:38we're doing heavy processing in the 0:39backend so that is one of the very very 0:45key points to our success as far as the 0:48scale that we have to deal with by kind 0:51of standardizing our development process 0:53and standardizing our micro services in 0:55that regard and made the deployment 0:57secure Bernays and the monitoring and 0:58the metrics gathering extremely easy to 1:00do and it made most things just 1:02plug-and-play this do as a service mesh 1:04that runs the psychic renée's and it 1:06provides us with a standardized way 1:09inside the mesh to route traffic in 1:12between services and also allows you to 1:15block traffic that you wouldn't want 1:17coming into the network it also logs 1:20these erroneous events so if you do get 1:23like a DDoS or if you do get a an attack 1:27of some kind and you can kind of trace 1:29back and see what what's going on and 1:31with the system at that point in time 1:32ideally it would take some 100 1:34milliseconds to serve an ad and that's 1:37what we're able to achieve with we are 1:39able to distribute our system across the 1:41world using the cloud environment 1:46you 1:48[Music]