Graph databases will no longer be a luxury but will become a “must-have” for enterprise IT organizations in 2022, according to graph database provider TigerGraph.
According to Gartner’s research, by 2025, graph technologies will be used in 80% of new data and analytics systems, up from 10% in 2021, facilitating rapid decision-making across the enterprise.
Scalable graph technology can be a key link between massive amounts of data and key business insights, said Todd Blaschka, COO at TigerGraph in an interview with The New Stack. Indeed, graph technology will become a key competitive advantage for companies in various industries — from financial services and healthcare to retail and manufacturing.
Moreover, many companies are using graph technology to quickly highlight, discover, and predict complex relationships within data — insights that uncover financial fraud or help solve logistics challenges within the supply chain, Blaschka noted at the company’s Graph + AI Summit in New York in October.
For instance, based on insights gained via graph technology, Mastercard has developed an algorithm to detect credit card numbers for sale on the Dark Web, based on the last four digits of the card, to zero in on compromised merchants, said Rohit Chauhan, EVP of AI and Security Solutions at Mastercard, who spoke at that same event.
Applying the Power of Graph Analytics
“Throughout 2022, more companies will apply the power of graph analytics to support advanced analytics and machine learning applications, including fraud detection, anti-money laundering (AML), entity resolution, customer 360, recommendations, knowledge graph, cybersecurity, supply chain, IoT, and network analysis,” said Dr. Jay Yu, VP of Product & Innovation at TigerGraph, in an email exchange with The New Stack and in a separate blog post. “Graph will become even more linked with ML and AI. Gartner even reports that ‘as many as 50% of Gartner client inquiries around the topic of AI involve a discussion around the use of graph technology.'”
Moreover, TigerGraph is working to advance graph databases, and has committed to delivering petabyte-scale graph databases within the next two years, Yu said.
The company is also working on a 100 TB version of the LDBC-SNB benchmark (Linked Data Benchmark Council Social Network Benchmark), which should be achieved within this year. In terms of standardization, the next year will see a move towards finalizing standardization on Graph Query Language (GQL).
“We also anticipate a deployment shift from on-prem to cloud (both private and public),” he said.
The Rise of Graph Data Science in 2022
In addition, Yu predicts a rise in graph data science in the coming year.
“We will see the rapid adoption of graph data science, as data and learning are naturally merged,” he said. “Relationships are fundamental, so having meaningful patterns is critical to enhance conventional machine learning and provide explainable results in industries such as healthcare (to provide real-time recommendations), the industrial supply chain (leveraging analytics for decisions), and financial services (to power real-time fraud detection). In fact, we will see the fusion of distributed graph databases, advanced analytics, and in-database machine learning.”
In addition, graph data science brings expert knowledge to machine learning via graph analytics, and more enterprises will begin to use it in production over the next year, he said.
“These enterprises will benefit from deep-link pattern discovery, uncovering patterns and insights within 10-20 hops (or connections) of data,” Yu added.
The Rise of Digital Twins
Meanwhile, Richard Henderson, Technical Director at TigerGraph, predicts a slew of new graph analytics industry use cases and vertical solutions such as “digital twins” will crop up and they will be based on real-time analytic graph databases.
Borrowed from the manufacturing industry, a digital twin represents a real-time model of a business and its environment.
“A digital twin gives a complete and current view of the physical business situation, using the business events and data that are probably already available in individual operational silos and data marts,” Henderson wrote in the post. “This, combined with graph analytics, can deliver a detailed and immediate digital scenario, showing the impact and risk associated with any delays or failures within that network. Graph analytics will combine these individual events and the map of the network to produce a completely ‘zoomable’ big picture view of the entire operation, so decisions can be prioritized based on both tactical and global concerns — with full knowledge of the impact of specific decisions.”
Using digital twins, an organization can do a what-if analysis for any changes or alternate scenarios to help avoid unintended consequences and reduce risk.
“The beauty of ‘analytic graph databases’ is that they are generic, adaptable, and powerful enough to express digital twins in a simple, direct, and rapid way,” Henderson said. “This means the technology can be quickly applied in whatever domain it may provide benefit. The result: the joined-up viewpoint that businesses have been asking for that will drive resilience in the form of real-time, data-driven operations.”
TigerGraph Pros and Cons
TigerGraph continues to advance the graph market while working to democratize graph adoption, Blaschka said. Last year, the company raised $105 million in Series C funding, one of the largest funding rounds within the graph database and analytics market, bringing TigerGraph’s total funding raised to over $170 million. Also over the last year, the company more than tripled its workforce and has expanded its global footprint throughout North America, EMEA, and APAC.
“They are a respectable contender in the graph database space,” said Andrew Brust, CEO of Blue Badge Insights.
In a recent report, George Anadiotis, founder of Linked Data Orchestration and an analyst at GigaOm, listed among TigerGraph’s strengths: Advanced analytics support; out-of-the-box graph algorithms; built-in IDE and visual exploration; third-party integrations; good onboarding and many options for data import/export.
However, Anadiotis, who also presented virtually at the company’s Graph + AI Summit, acknowledged that TigerGraph has its share of challenges. Among them: “Knowledge graph use cases are supported, but lack the ease of use and optimizations RDF [Resource Description Framework] vendors provide; limited options for on-premises deployment; a relatively small community,” he listed.
TigerGraph’s built-in IDE, GraphStudio, offers advanced capabilities such as visual query construction, Andadiotis said.
“GraphStudio is like Visual Studio for graph,” Yu told The New Stack. “It’s visual because graph is visual.”
Meanwhile, TigerGraph also works with third-party IDEs and data science notebooks. Visual exploration can take place either via GraphStudio, or via third-party integrations, including BI tools, Anadiotis noted.
The latest version of TigerGraph is release 3.2, which the company introduced in October with 40 enterprise enhancements aimed at boosting performance and scalability, simplifying management of the database and increasing security.
TigerGraph 3.2 brought simplified management via cluster resizing, faster backup and restore, and direct control over resource allocation for big queries, Yu said in a blog post announcing the release. It also featured cloud management via built-in Kubernetes support, and security and access control at scale via user-defined roles.
Moreover, version 3.2 added developer enhancements to make graph technology accessible to more organizations, and to enable business users and data scientists to go deeper into data without having to write code. New features included added accessibility compliance, query language enhancement, and query build performance speedup.
Meanwhile, TigerGraph is swimming in a pool of graph competitors including the more established Neo4j and offerings from the likes of Amazon, IBM and Oracle to name a few.
For its part, Neo4j ranks as number 20 on the DB-Engines Ranking of database management systems according to their popularity. It is the first graph database ranked. TigerGraph is ranked 100 slots below at number 120.
“My sense is that this category is growing in efficacy and that Neo is the far-and-away leader,” Brust said. “It’s sort of like there’s Neo and then there’s everyone else.”
Neo4j has a growing developer base. According to the company, the global Neo4j community surpassed 240,000 members over the last year. In 2021, developers downloaded Neo4j more than 36 million times and launched more than 150,000 Neo4j Sandbox instances. Also, upwards of 53,000 professionals list Neo4j as a skill on their LinkedIn profiles, the company said.
Featured image via Pixabay.