[GoSF] EVCache, Peripheral I/O, Building Origin Cache for Images

We are collaborating with GoSF meetup group to stream their meetups live through our channel.

• Talk 1: EVCache: Netflix's High-volume, Low-latency Caching Service by Scott Mansfield

This talk gives an overview of the EVCache system at Netflix, including high-level features, use cases, and architecture. The second part of the talk is about the next generation server that is currently being rolled out internally. This piece of the project hinges on its use of Go. It includes a couple new components: Rend (https://github.com/netflix/rend) which is a high-performance memcached compatible proxy and server, and Mnemonic, which is our disk-based storage process.

Bio:

Scott Mansfield is a Senior Software Engineer at Netflix. He works on EVCache, a high-performance, low latency persistence system. He is the primary author of Rend, an open-source memcached proxy as a part of EVCache.

• Talk 2: A proposal: Peripheral I/O for Go by Jaana Burcu Dogan

Do you include Go in your hardware projects? So do we. Over the past years, there is a significant portion of the Go community used Go to write hardware controllers. There are multiple successful and popular libraries yet no consolidation on the core interfaces. We are proposing a group of peripheral communication protocol packages for the Go programming language. Portability, extendability and reusability are the major goals. In the scope of this talk, we are going through the motivation, current state and the future goals of the project.

Bio:

Jaana Burcu Dogan is a engineer working on embedded systems at Google and is a Go contributor

• Talk 3: Building an Origin Cache for Images by Cindy Sridharan

At Imgix, our multitiered caching layer forms an important components of our real­time image processing and delivery service. This talk will shed light on what necessitated the system and the architecture, how several of the language’s inbuilt constructs as the tooling were used to design, implement and test the cache filling, eviction and purging logic, and more interestingly, our experience ‘productionizing this’, tuning the system to handle disparate traffic patterns, the mistakes made, challenges faced, lessons learned building this and what the future holds for this system.

Bio:

Cindy Sridharan learned Go in the fall of 2015 and currently works at Imgix where Go is used to build components of the origin cache. She organizes the Python Twisted and Lua meetups in San Francisco.

8:30

End