Ontotext

GraphDB includes various capabilities that extend beyond the database, notably ontology visualisation and connectors to a variety of third-party environments (mostly search engines such as Solr, Lucene and ElasticSearch, but also MongoDB). Internally, GraphDB’s inference engine employs forward chaining, and it provides various functions for dealing with materialised inferences which retract inferred statements. Cluster management and performance (including load performance), security (including LDAP and Kerberos support), visualisation capabilities and support for workflow processes to enable the loading of structured data into GraphDB are all major features, as is support for geo-spatial constraints and graph-path search. In addition, GraphDB offers several features for delivering high-availability clusters, including quorum-based replication and node-driven, self-organising leader election.

GraphDB supports RDF* and SPARQL*, which allows you to attach attributes to relationships. This is the equivalent of labels on a property graph, except that RDF* is more expressive. In addition, the company’s implementation of RDF* supports backwards compatibility.

What’s more, Ontotext is intent on making graph applications and analytics really easy to use for developers. For instance, GraphDB supports GraphQL via the Semantic Objects add-on, a framework, that allows the relevant APIs to be automatically created from semantic object descriptions, essentially meaning that developers do not need to know SPARQL. It also provides built-in facilities for data import and reconciliation (now provided as the separate Ontotext Refine tool), querying, search, visualisation, monitoring and cluster management, plus the company focuses on working and integrating with various open-source components. This should all help to provide an easy start for developers.

Ontotext has also developed deep integration with many of its partners in order to deliver best-of-breed knowledge graph management. These partners include PoolParty, Metaphactory, Synaptica, TopQuadrant, Eccenca Corporate Memory, ONLIM, Perfect Memory, and Squirro, among others. This means that GraphDB acts as a gateway to various complementary technologies, such as ontology and taxonomy management, data catalogues, enterprise search, and so on.

Ontotext impresses with its dependable performance and efficient handling of both graph analytics and semantic metadata management. This is borne out through its audited results in both the Semantic Publishing and Social Network LDBC benchmarks.

In addition, Ontotext has the best implementation of RDF* that we have seen, and this combines with the company’s performance on the aforementioned benchmarks (especially the Social Network Benchmark, which is designed to measure graph analytics capability) to give real credence to the idea of using Ontotext for use cases that have traditionally been reserved for property graphs.

The scope of the Ontotext Platform is broader than most anything else in the market (in part thanks to its long list of integrated partners), and we like its support for GraphQL and its integration with open-source technologies of various kinds.

Moreover, GraphDB’s support for generative AI, LLMs and ChatGPT puts it noticeably ahead of the curve when it comes to these areas, which are currently some of the most popular in data.

The Bottom Line

Apart from its domain expertise, Ontotext has historically offered significant capabilities in semantics, search, text mining, and knowledge graphs. Indeed, its continuing emphasis on the latter is both notable and sensible, as it is one of the most sought-after graph use cases. In short, if you have an interest in knowledge graphs, and especially if you have an interest in using them alongside generative AI, you should consider Ontotext.