SingleStore

Checked - One solution for Streaming Analytics on Big Data

Puntos a favor:

- Super simple, we implemented the whole of our customer-facing streaming analytics data store in a 2-day hackathon - Natively integrates with AWS & other services (we are a multi-petabyte S3 shop with huge Kafka clusters) - Since most of it is SQL, anyone can hit the ground running in no-time, an ultra-low barrier to entry - We saw a sub-second response on a test case with 1000 concurrent heavy API calls (scanning billions of rows across many tables + window functions etc) along with applying RBAC functionalities on the fly - Incredibly fast out of the box, basic performance tuning will suffice 95% of use cases - Fantastic partners to work with, super helpful, we have access to their core engineering team, +100 on Customer engagement!

Contras:

In my humble opinion, it solved a very unique hard problem for us elegantly (without having to compromise on any architectural or data principles). Problem of streaming analytics on high-velocity high volume data sets with sub-second API responses can be a handful. For me, this along with other use-cases they mention is their core strength. So anything beyond these capabilities is a "Nice to have" for me. Their roadmap covers most of what the community is asking - hot back-up's, better UI, modelling etc.

Thoughts on MemSQL

Comentarios: Our primary use case for MemSQL is an enterprise data lake, which will also serve as a staging area for MemSQL data warehouses. They are currently being developed and are not in production yet. We're learning things as we go, but in general, development is going well and we're satisfied with the product.

Puntos a favor:

I appreciate that MemSQL is a distributed database that runs on commodity hardware and can be scaled out as opposed to some fixed, vendor-maintained appliances like Teradata. Columnstore is a valuable feature.

Contras:

The product is not fully mature. Although it uses the MySQL protocol, it does not always work with MySQL-compliant products. Some administrative functions such as incremental, point-in-time, and object-specific backups and restore are lacking. MemSQL is making progress in these areas, but it still has room to grow.

MemSQL : Needs Time to Mature

Comentarios: This is a good software but needs to mature more for an enterprise level

Puntos a favor:

Similar to existing RDBMS Supports join In Box implementation with Spark Salable & manageable with memSQL ops

Contras:

Security Features Consistency Overflow on Disk

SUPERSONIC DB - SUPERB

Comentarios: I'll rate it at a full star. Having worked with almost all the databases, I can simply say it aces every database that I worked. The Transformation capabilities in LOAD command is something I have not seen anywhere which helps a lot along with capturing errors and rejects in the LOAD command.

Puntos a favor:

Super fast and would have give a fair bit of competition to CONCORD Aircraft:). But seriously it is the Database of the new generation with huge potential for both On Premise and Cloud. High compatibility, resilience, scalability are some superb features to work on with it. Highly recommended to any Organization wanting to get rid of the Old fashioned drag on DBs.

Contras:

I don't have any particular dislikes, but would love to see its own Client interface, inbuilt Data Modelling capabilities and a superb front end which can help potentially any Developers or Programmers to start the development from Day 1. Secondly I would also like to see a proper Blogging web portal to help developers resolve their queries, issues faster.

Very fast

Puntos a favor:

fast response, perfect for real time apps

Contras:

documentation isnot too clear about what sits in memory and what on disk. even though all our tables were made to reside on disk, we were getting lots of usage alerts, took for ever to resolve.