It has been four weeks since the ScyllaDB Summit and we’re still excited about it. The Summit was a big success and we had a packed room for our early workshop session – turns out a workshop right after Labor Day isn’t a bad idea after all!
Let’s have a look at ScyllaDB Summit by the numbers:
- 1-day event, 100+ attendees
- 6 partner/customer presentations
- 5 Team-ScyllaDB presentations
- 50 ScyllaDB clusters created at the workshop
- 2 ScyllaDB DBaaS announced
The excitement around ScyllaDB continues to grow and our modest db-engines ranking spiked up 30 locations. We have a long way to the top, but that’s our goal. A few years ago, the Cassandra Summit was a small event as was mentioned at their keynote and demonstrated in a C* community growth graph, here::
Think about where ScyllaDB will be in three years. ScyllaDB is a young two-year old project, however, all the ecosystem around Cassandra is preserved and we plan on taking it to the next level while embracing
both the Cassandra community and our own growing community.
Key presentations highlights at ScyllaDB Summit
Keynote – Big Data Goes Native
ScyllaDB isn’t just a faster Cassandra. By revoking the underlying implementation, we unlock the power that allows consistent performance over time, while running mix workloads and executing administrative tasks such as repair, compaction and streaming. The ScyllaDB project strives to be the next generation platform for a variety of workloads, from time series, to graph and analytics.
Outbrain Case Study
Outbrain is the largest recommendation engine in the world. ScyllaDB allows Outbrain to migrate away from a Cassandra plus Memcache setup which is spiky by nature, suffers from cache inconsistencies and cold-cache issues. This session demonstrates that ScyllaDB handled 20X the amount of traffic than Cassandra while providing 3X less latency.
ScyllaDB and Cassandra at Kenshoo
Kenshoo is the industry leader in digital marketing and has run Cassandra for quite some time now. The Kenshoo team was interested in ScyllaDB not just due to performance but for ease of administration. It is important to be able to run all of the background operations a sophisticated database keeps under the hood. Kenshoo engineer Noam Hasson demonstrates the speed of compaction and repair of ScyllaDB verse Cassandra. Eventually Noam shows that three ScyllaDB VMs can handle the load of 40 physical Cassandra machines.
IBM’s Compose: Database in Containers
Compose, an IBM company, announced the launch of ScyllaDB as a service on their platform.
ScyllaDB was chosen on top of Cassandra, mainly due the ease of configuration and tuning. No JVM needs to be tweaked. Performance is great out of the box with the good old compatible CQL verbs. In this talk, IBM Compose’s Nick Stott explains the how and why of simplicity and resilience for databases.
IBM Graph: Graph Processing with Titan and ScyllaDB
Jason Plurad with IBM graph and Apache TinkerPop committer describe how TinkerPop, Titan and ScyllaDB are a match born in heaven.
ScyllaDB in Samsung SDS
Dror Gadot and Kuyul Noh of Samsung SDS announced that they are joining the project and shared their own PoC, which shows gains between 2-8X of throughput along with additional latency gains. The session covers Samsung SDS’s journey from ScyllaDB 0.10 to 1.23, from functional to stability test. You’ll hear about the business use cases for which Samsung SDS is using ScyllaDB, such as IoT Data Platform and Messenger service, and learn what the future holds for Samsung SDS’s contribution to ScyllaDB community and business partnership with ScyllaDB.
ScyllaDB – Present and Future
Avi Kivity, co-founder and CTO revealed our upcoming roadmap. Soon we’ll achieve parity with Cassandra with features such as Secondary indexes, counters and lightweight transactions. In parallel, we’re heavily invested in #workloadconditioning. The later is a family of algorithms, from the cpu to the network which transparently adjusts the run time parameters and behavior of ScyllaDB to cope with all kinds of workloads.
Try it yourself
Isn’t it time you’ll take ScyllaDB for a spin? There is a Docker image just for that!