Our great sponsors
-
-
InfluxDB
Power Real-Time Data Analytics at Scale. Get real-time insights from all types of time series data with InfluxDB. Ingest, query, and analyze billions of data points in real-time with unbounded cardinality.
Someone from ScyllaDB here (Technical Marketing Manager). Forgive, if you can, the length of my reply.
AGPL was chosen to prevent people from taking the software and making it an -as-a-Service (-aaS) offering without contributing anything back to it. Which, if you look at other open source products, can cause them to wither in the vine as people reap the benefits without having to sustain and enhance the base code.
We now have plenty of folks using our Scylla Open Source product across spaces from cybersecurity to IIoT. No one who is just using Scylla internally really needs to worry about AGPL. Though I do admit that many people are allergic to it for lawyerly reasons. But it's also helped prevent other not-so-fine people from utterly vulching the code.
Scylla Open Source is often used under JanusGraph, which is the open source fork of TitanGraph now supported by the CNCF (folks familiar with the history know what happened to TitanGraph, so yes, your concerns are warranted). We use open source Prometheus and Grafana for our monitoring, rather than a proprietary offerings.
We're also taking your first point seriously (long-term direction). We see ourselves as stewards of the software; we don't want to bottleneck or freeze out contributions. For example, open source contributor @Fastio began adding the Redis API into Scylla Open Source! I remember when I learned he was planning on doing it, beginning with a Redis on Seastar implementation called "Pedis." Now it's there in the open source code base. Pretty amazing work, and you have to just thank amazing contributors like that.
https://github.com/scylladb/scylla/blob/master/docs/design-n...
https://github.com/scylladb/scylla/tree/master/redis
Apache Cassandra is also an awesome project, and ScyllaDB definitely owes a lot of our success to the groundbreaking work done there. Anyone working on it gets nothing but big props from me.
We therefore also want to ensure that what we do stays pretty much compatible with Cassandra (CQL v4, murmur3). Like the new Rust driver we wrote as part of our internal hackathon:
https://www.scylladb.com/2021/02/17/scylla-developer-hackath...
While the rivalry with the Cassandra community remains pretty heated in some parts with some parties, you'll get none of that from me. Personally I just hope that end user developers just get better code, better features, better choices.
In 2018, the head-to-head rivalry seemed pretty fierce. But now there are soooo many closed source CQL offerings out there: DataStax, Amazon Keyspaces, Azure CosmosDB, Scylla Enterprise (separate from our open source). There's also other open source offerings like Scylla Open Source and Yugabyte. Of all of those, we hope to show up as the "most open" of the competing offerings.
Also as of 2021 Scylla has broadened who we can please (or, I suppose, be mad at us) by offering other APIs. We support a CQL interface for Cassandra compatibility, a DynamoDB-compatible API, and, still under development, the aforementioned Redis API.
Each of those different NoSQL communities and constituencies bring high expectations for excellence, and their own high standards for what they want from an open source vendor. We definitely take their criticisms to heart.
And yes, our DynamoDB implementation, Alternator, is fully 100% open source. You can totally run your workloads where you want. On premise, on any cloud, or even still on AWS. We take that aspect of open source very seriously. We could have made it simply an enterprise feature. But we opened it up.
I know my title is "Marketing" and some people see that as a license to lie on behalf of a vendor, but I have never been more proud to see the open source commitment and contributions of any company I've worked for to date.
Thanks for the mention and for reading this far. And best wishes to anyone working on hard big data problems these days, regardless of your database-of-choice.