Re:Anna: A Fast, Scalable, Flexibly Consistent Key-Value Store
by zzzcpan on Friday March 09, 2018 @03:10PM
(#16554579)
Key-value store is not the focus of this research. And the problems it focuses on are not even in the same league as those you are interested in.
Re:Anna: A Fast, Scalable, Flexibly Consistent Key-Value Store
by snissn on Friday March 09, 2018 @02:11PM
(#16554065)
> Why so much focus on Key Value stores? That's the easy part of the problem.
There aren't enough good/fast/reliable ones
Re:Anna: A Fast, Scalable, Flexibly Consistent Key-Value Store
by drb91 on Friday March 09, 2018 @03:01PM
(#16554503)
Is that a CAP joke or what?
Re:Anna: A Fast, Scalable, Flexibly Consistent Key-Value Store
by kamranjon on Friday March 09, 2018 @06:16PM
(#16555919)
Does Redis not count?
Re:Anna: A Fast, Scalable, Flexibly Consistent Key-Value Store
by doorbumper on Sunday March 11, 2018 @01:55AM
(#16561976)
Redis is an impressive piece of engineering, but it performs best as an in-memory kv-store on a single core. Its distributed capabilities target a different problem than other distributed kv-stores attempt to solve. Redis Cluster focuses on reasonable functionality for an in-memory store. However, Redis Cluster is neither highly available, nor consistent. There are multiple modes that can cause catastrophic data loss, so Redis Cluster works best in situations where losing data isn't a big deal. For it's intended use case, nothing else comes close to offering the same functionality, performance, reliability, and ease of use.
Re:Anna: A Fast, Scalable, Flexibly Consistent Key-Value Store
by stochastic_monk on Friday March 09, 2018 @12:17PM
(#16553182)
R-trees (and their higher-dimensional counterparts) and B/B+-trees are great for supporting range queries. You’re right that kvs systems are limited, and I find the above quite effective at supporting more general operations.
Re:Anna: A Fast, Scalable, Flexibly Consistent Key-Value Store
by zilchers on Saturday March 10, 2018 @12:58PM
(#16559421)
CockroachDB is implemented on top of a key / value store. I don’t know too much about the internals, but a kv store with lots of degrees of consistency seems like a good primitive to build more complex data stores on top. For instance, indexes could live as immediately consistent sharded keys, while data stays eventually consistent (or user configurable).
Re:Anna: A Fast, Scalable, Flexibly Consistent Key-Value Store
by majidazimi on Friday March 09, 2018 @02:45PM
(#16554362)
Here is my 2 cent on why this is the case:
On a general purpose hardware/network adding features as you mentioned, tends to be seriously difficult and tremendously workload dependent. At some point you need help from hardware designers which itself limits the use cases of your system.
FPGA supported databases, In-network SQL processing, Server-less caches (an SSD directly connected to network without OS), Storage systems with mind blowing low energy consumption, ... are just bunch of absolutely amazing research that are being done today. however the limited public availability is badgering.
With mass public usability in-mind, a general purpose K/V system is the highest summit you can really achieve.
Why so much focus on Key Value stores? That's the easy part of the problem.
I would like to know more about the interesting ones: secondary indexes, range scans, performance on mixed workloads, robustness, operational complexity.