2018 in Review

Time to look at our 2018 progress. What’s the state of Stardog for 2019?

Looking Back

It feels like 2018 was the year that Knowledge Graph started to break through the buzz and really get into the imagination and strategic plans of the Fortune 1000, which is our primary focus. I’ll just mention three anecdotal sorts of proof that have us excited here.

First, we started talking about Stardog as the Enterprise Knowledge Graph platform in late 2016, early 2017. By the second half of 2018 it wasn’t just us doing the talking, but enterprises in financial services, manufacturing, government, and health care life sciences were telling us that a Knowledge Graph–to unify data, to enable analytics and insight, to accelerate digital transformation–was on their critical roadmap to learn more about, to evaluate, and to put in production.

Second, vendors of plain graph database systems, a kind of data management tool that we’re quite fond of, have fully glommed onto the Knowledge Graph idea as valuable and momentous. You know how it goes. First they ignore you, then they make fun of you, and then they steal your messaging and positioning. But we say the more, the merrier. This tide will raise all of our boats.

Finally, in the analyst space, Gartner recognized Knowledge Graph as an emerging technology and recognized Stardog’s semantics-based approach as strategically important. One of the ways I spend my time nowadays is telling the Stardog story to industry analysts and learning to listen to their feedback as an important buy-side market signal.

We think all of these are great signs that 2019 will see even greater market penetration, not only for the idea of a Knowledge Graph, but also for the reality of Stardog as the most mature and innovative offering in the space. No other system can touch Stardog in terms of capabilities in service of a complete vision and we’re not done yet, not by a long shot.

Product

So what did 2018 look like in terms of product? A few software engineering and release stats should get us started:

• 5.2.0 was the first release of 2018 (on January 10th)

• We did 16 releases in 2018. Which is weird, since we release monthly. In those 16 releases,

  • We fixed 186 bugs
  • We made 129 mods (changes other than bug fixes or major new features)

• 7.0.0-beta1 was the last release of 2018 (on December 19th) and that one almost killed the storage engine team.

Since it’s 5:30am on the day before Christmas, and I have a long drive to make yet today to get presents for my kids, I’ll just list the major new capabilities and services we added in 2018 to the Stardog platform:

1. Named entity recognition and linking
2. Similarity search
3. GraphQL introspection, extend GraphQL to Stardog search
4. Automatic indexing & mapping of any RDBMS databases
5. Virtual graph support for MongoDB, Apache Hive, Sybase ASE, IBM DB2, Apache Impala, Amazon Aurora, Cosmos DB
6. New Virtual Graph mapping language, SMS2
7. Kerberos support
8. Machine learning with XGBoost
9. Customizable DESCRIBE query strategies
10. S3 backup & restore
11. Stark, new core Java API
12. Add built-in JTS support to geospatial
13. Massive HA cluster stability and throughput increases
14. Signficant concurrent write performance increases
15. New backend based on RocksDB, LSM Trees, and LCW conflict semantics

When you write it all down in one place like that…I think Stardog engineering is the best in the biz.

Stardog Studio, Too!

And lest I forget the front-end team, 2018 saw the inaugural release of Stardog Studio, in February, and the Studio team made 20 releases in total this year. That team ships product like they mean it!

Since all of Studio is new in 2018, I’ll only list some of the major improvements after the initial release in February:

1. Display query plan
2. Support for stored queries
3. Query management console
4. SPARQL completion, heuristics, hints, and full-on IDE support
5. Full IDE support for graph data language (Turtle)
6. Virtual Graph mapping support, including full SMS2 support
7. GraphQL support

Studio is a vital part of the Stardog platform and you should expect 2019 to pick up the pace as we’re expanding that team with two new developers.

Top Blog Posts

This is our 37th blog post of the year, which is about what we did in 2017 (38), but fewer than we intended. For 2019 we intend to average one new post per week. Get hyped, friends.

The top 6 posts on this blog in 2018, in no particular order were largely clustered around practical and applied themes, which makes sense given the maturity of Knowledge Graph in 2018: entity linking, geospatial queries, even more about entity linking, anti-money laundering, customizing Stardog’s NLP, and unifying MongoDB and other semi-structured data silos.

Team

2018 also saw changes in Stardog personnel, including a new CEO, Michael Sachse, and a new Chief Scientist, Dr Evren Sirin, who, of course, isn’t new at all, being one of the Stardog founders, alongside the inimitable Mike Grove. And just recently we added a new, world-class independent director to our Board, Dan Yates, who was one of the founders of Opower. We also added many new members of the go-to-market team, including VP Marketing, Lauren Taylor, and new engineers in all platform subsystems.

Looking Ahead

Retrospectives are, of course, a fictional genre. I’m focusing here on the stuff that worked well. As with any growing, healthy startup, we had plenty of false starts, failed experiments, and good old-fashioned mistakes. I will keep the details to myself and only say that we have a lot more to learn.

Product

In terms of the core platform and Studio, we’ve been sharing roadmap with customers and other interested parties and will continue to do so. You should join one of our upcoming webinars for details. At a high level, I can promise:

• Studio will continue to evolve into the Knowledge Graph IDE, including more query performance diagnosis tools and smart completion across the Knowledge Graph
• more Virtual Graph support for the world’s most crucial data management systems
• bigger clusters as the core platform moves to Kubernetes
• Kubernetes-scaled core services including Virtual Graphs, search, reasoning and machine learning
• GA release of the new, native backend with crazy concurrent write speedups

A little bird has been whispering to a dog I know about some new Stardog products in 2019, including stuff about clouds and clusters. I’m not sure, but it seems quite promising.

Team

We’re hiring. I count 12 open positions on the Stardog Careers page. Come help us unify all the things!

Conclusion

We made real strides in 2018 and have lots of ideas, plans, and a few surprises for 2019. I’m proud of the hard work, devotion to craft, and sacrifice my coworkers make every day. Even more so, however, I’m excited to show you all the amazing things they will build in 2019. In the mean time, why not join an upcoming webinar to learn what all this Knowledge Graph stuff is all about? Happy Holidays and Happy New Year!

Download Stardog today to start your free evaluation.

• Company news