Implement publish, subscribe, blpop, brpop

[sscarduzio]

This stuff is pretty important part of Redis, people will have difficulties adopting your technology if migrating from Redis to SSDB involves giving up features.
You went so far doing a great job with this project, why not to have these implemented?

[ideawu]

It is better to get these features by using message queue service like ZeroMQ, Gearman, etc...

[sscarduzio]

Generally speaking I agree with you, but I have a few objections:

• Not all the projects are web-scale (or are planned as such), a startup might decide to keep the technological stack simple and use Redis for data, queue broker and pub-sub. Accidentally, these are also the same startups that would benefit from the cost reduction effectiveness of not keeping all data in memory.
  Remember also that startups tend to contribute quite much to OSS and innovate more easily.
• Probably also at high scale, an internal, low traffic Redis publish-subscribe channel is a good idea for coordinating services, without the need of setting up ZMQ for a simple heartbeat service.
• You'll be amazed, but Redis makes a hell of a good queue broker (not all projects require or benefit particularly by brokerless queues). And there's a plenty of projects implementing queues in redis.
• PubSub in Redis is really fast too (although it naturally suffers with a very high number of subscribers or, dually, a very high subscribe traffic).

[Kortenbach]

I've only recently found out about SSDB. Great work!
Since SSDB is made as an alternative to Redis I would expect it to include pub-sub. Because Microsoft dropped Redis support I was looking at SSDB as an alternative, but I can't use it because it doesn't support pub-sub.
One thing I really missed in Redis pub-sub messaging is the key values (old-new). I would really appreciate if that would be added to SSDB!

[gokaybiz]

It seems like this project abandoned 🫤

[alanhamlett]

It seems like this project abandoned 🫤

Yep, I've switched to https://github.com/apache/kvrocks