Considerations for multi-data-center applications
Most applications should at least consider running in multiple data centers for reasons from end-user latency to being able to meet the modern expectation for 24/7 uptime. Jeff Poole outlines the considerations when moving to multiple data centers and the trade-offs for different approaches.
| Talk Title | Considerations for multi-data-center applications |
| Speakers | Jeff Poole (Vivint Smart Home) |
| Conference | O’Reilly Software Architecture Conference |
| Conf Tag | Engineering the Future of Software |
| Location | New York, New York |
| Date | April 3-5, 2017 |
| URL | Talk Page |
| Slides | Talk Slides |
| Video | |
Jeff Poole offers an overview of the reasons why you might design a system to run in multiple data centers. You get performance improvements: improving end-user latency and transfer speed as well as reducing lag in real-time media and command-and-control applications. More importantly, being geographically distributed can improve uptime and resilience to natural disasters or network issues that could plague a single data center. Jeff then explores the concepts to consider when moving to a multi-data-center deployment: Jeff concludes by walking you through some scenarios of different applications, explaining how trade-offs may be applied to those scenarios and discussing the different choices you might make between a real-time communications platform and a data-heavy web application, and how those choices might be moderated by different uptime requirements.
