I’m here at RedisConf in San Francisco — I’ve been wanting to go for a couple of years now but it’s always been a scheduling problem on my end. But despite moving all last week, I was able to carve out enough time to make the trip.
Day 1 Keynotes Highlights:
- Salvatore offered some interesting glimpses into the future of Redis. 4.0 is big, but I’ve known about it for a while. The streams data type in Redis will be a huge boon for time series stuff.
- Kelsey Hightower did a damned impressive demo of Redis Enterprise Cluster running with Kubernetes. His showmanship in the demo had impeccable timing. He was also administering it with voice command. It’s hard to describe, but I think everyone was wowed.
Recap on the sessions I attended:
Session 1: Redis Open Source vNext, Redis Stream
A little deeper dive with Salvatore again. Quite interesting to see the threading complications that 4.0 had to face and how the radix tree and listpack will change performance and memory efficiency. The future is bright.
Session 2: Preventing Cache Stampede with Redis and Xfetch
Jim Nelson from Archive.org described how caching can get complicated and how to mitigate it with the Xfetch algorithm. Xfetch was new to me, but fascinating. It’s a probabilistic way to efficiently know when to re-fetch (recompute) something stored in cache. He implemented it in php, but I think I might toy with it in Node.js soon.
Session 3: Ice wine, Redis and Beyond: Redis on a Microcontroller
My session! It went pretty well. I wish I had been able to actually demo the microcontroller running as a Redis client, but the WiFi in the room required a login prompt so the microcontroller wouldn’t have been able to connect. The monitor/projector setup was also a little difficult manage physically, otherwise I would have shown more code. Despite these little blips, it was a lot of fun.
There was some definite interest in the room about how Redis and IoT/Microcontrollers can go main stream. I’m looking forward to building a community around it.
Session 4: Deploying the RediSearch Module at Scale and an Intor to the ReJSON Module
Itamar subbed in for Dvir on the RediSearch module. The RediSearch module is something I’ll be writing about soon, so this was vital to me. There were a few things that clicked beyond what I had read already on the module. The indexing speed demo was impressive. ReJSON was also covered, although RediSearch was the star of the show.
Session 5: Distributed Business Rules using Redis
Jesus Ruiz presented on how business rules engines can take advantage of using Redis. I have to admit that I was a little clueless on business rules engines, but I can definitely see how they would be useful. It was also really interesting to see how, by using Redis, he was able to reduce the time of computation from 4.7 seconds to 600 microseconds.
Can’t wait to see what day 2 at RedisConf holds!
Update! Here is my recap of day 2 and conclusions about RedisConf.