Frequently Asked Questions (FAQ)

How do you pronounce the name Dydra?

It’s pronounced ‘die-dra’, rhyming with ‘hydra’. (‘dei-dra’, auf Deutsch.)

Who is behind Dydra, and where are you based?

Dydra is a product of Datagraph, Inc., a privately-held U.S. corporation headquartered in Delaware. Our development team is located in Berlin.

The Dydra website rocks, who designed it?

Our brand and website was designed with meticulous attention to detail by the very talented folks at Makalu Interactive.

How much would it cost to host X billion triples?

We continue to offer limited-resource subscription-free accounts, but if you need a host for a billion triples, please get in touch and we will explain our pricing details.

Do you support feature X, Y, or Z?

We have a baseline set of features which we make available for evaluation, and exciting additions which we make available for production users. We prioritize offering what our customers need, so please do let us know about your use case(s) for Dydra and which features are essential to you.

What are the configuration limits for an account on the public service?

  • Federation is disabled.
  • The default graph in the default data-set is fixed as just the default graph from the repository.
  • Undefined variables trigger an error.
  • Non-standard vocabulary terms trigger an error.

What are the service limits for the public service?

We enforce various soft and hard resource limits for evaluation accounts to help with our capacity planning:

  • As a baseline, we currently rate limit evaluation users to 5 queries per second. If you need a higher burst rate, please contact support@dydra.com.
  • We ask that you keep repository sizes at or below 1 million statements. Again, if you need a higher limit, please contact us.
  • This site is not intended to serve as a benchmark endpoint. There are so many variables to possible deployments that you will fare better, to propose your circumstances and provide us an opportunity to respond.
  • Your use must remain equitable and purposeful. Again, no benchmarking and requests with queries which repeat exhaustive scans, generate multi-megabyte result documents, or trigger time and other resource usage limits do not serve to convince of a useful purpose.

What are the quality of service goals for the public service?

We aim to provide continuous service, save for a brief weekly scheduled maintenance window on Saturday mornings at 08:00 UTC when 5-15 minutes of downtime can be expected.

There may be occasional brief unscheduled downtime now and then to address more urgent matters or to deploy additional capacity. Follow @dydradata on Twitter to get announcements.

What technologies is Dydra built on?

We’ll share more details in upcoming blog posts, but the website is built with nginx, Ruby and Ruby on Rails, relying heavily on the RDF.rb framework. Our back-end infrastructure is based on core messaging technologies such as AMQP and Redis. Our proprietary distributed SPARQL query engine, called SPOCQ, is built in Lisp and compiles SPARQL queries to native x86 machine code. Our distributed RDF storage system is built in C, relying on Raptor at the interfaces. Our stack as a whole is hosted on Amazon Web Services.