TigerGraph launches Savanna to aid AI development

[Pelican Press 0]

This is the hidden content, please

• Sign In
• or
• Sign Up

TigerGraph launches Savanna to aid AI development

TigerGraph on Tuesday launched Savanna, a cloud-based graph database designed to quickly process large amounts of connected data to feed AI and analytics applications.

Now generally available, Savanna is the next evolution of TigerGraph’s cloud native database platform. Features, among others, include pre-configured kits for applications such as fraud detection and customer insights, optimized performance using parallel processing and connections to new data sources including Snowflake and Apache Iceberg.

Unlike relational databases that only enable data points to connect to one other data point at a time, graph technology enables data stored in a database to simultaneously connect with multiple other data points. By doing so, users can much more quickly and easily discover relationships between data in graph databases such as TigerGraph and Neo4j than in relational databases such as

This is the hidden content, please

• Sign In
• or
• Sign Up

SQL Server and Oracle Database.

Common uses for graph databases include fraud detection, supply chain management and social networking. Now, with enterprises increasing their investments in AI development, AI is another use case for graph technology given that AI models and applications require high volumes of relevant data to be accurate.

Because Savanna advances TigerGraph’s cloud-based capabilities, the offering is significant for the vendor’s users, according to Matt Aslett, an analyst at ISG’s Ventana Research.

“Savannah appears to be a step forward for TigerGraph’s cloud service with [its] new features,” he said.

Regarding Savanna’s potential to aid AI development, the platform will be of interest for supporters of graph technology, Aslett continued.

“Proponents of graph databases argue that the native storage of relationships means that graph databases are intrinsically more efficient than relational databases for applications and use cases that depend on identifying and navigating the connections between entities,” he said.

Based in Redwood City, Calif., TigerGraph is a graph database specialist. Like many vendors, TigerGraph has made AI part of its product development plans and in May unveiled a generative AI-powered assistant aimed at making the vendor’s platform easier to use.

Both the introduction of AI capabilities last year and the launch of Savanna come during a time of change for TigerGraph.

The vendor contracted significantly after the COVID-19 pandemic, dropping from 430 employees to about 130, former CEO Hamid Azzawe said in May. In addition, with Rajeev Shrivastava taking over as CEO in August, TigerGraph is under its fourth CEO in less than two years.

New capabilities

Enterprises are

This is the hidden content, please

• Sign In
• or
• Sign Up

in AI development now that the wide availability of generative AI tools have the potential to make workers smarter and more efficient.

At the core of the AI-powered tools they are developing is data, which provides AI with its intelligence.

However, discovering the right data to train an application is not simple. Tools such as graph databases and vector databases that enable similarity searches — which lead to the discovery of relevant data — have gained prominence. But to truly meet the needs of AI developers, such tools also need to perform at the scale AI development demands.

TigerGraph’s Savanna is designed to meet both the data discovery and scalability requirements of AI development, according to Shrivastava.

“Graph databases have always provided understanding and intelligence that other types of databases cannot because connected data is contextual data — explicit and explainable,” he said. “Savanna delivers a cloud-native graph database with the scalability and performance needed to serve today’s AI demands.”

Scalability and performance are enabled in part by parallel processing that provides dedicated compute workspaces for each online transaction processing (OLTP) and online analytical processing (OLAP) workload. In addition, Savanna’s ability to scale up or down to meet workload demands of any size plays a role in its efficiency.

Beyond better performance and scalability compared to TigerGraph’s previous capabilities, Savanna includes the following:

• Nine pre-configured kits for various applications.
• Cost savings from the separation of compute and storage under a pay-as-you-go pricing model.
• Connections to three times more data sources than in the past, including data lake and database sources Delta Lake, Iceberg, Postgres, Snowflake and Spark; TigerGraph already connected to AWS,
  This is the hidden content, please

  • Sign In
  • or
  • Sign Up
  Cloud and
  This is the hidden content, please

  • Sign In
  • or
  • Sign Up
  Azure object stores.
• Support for three graph query languages.
• A bring-your-own-cloud option so customers can choose between a fully managed service or deploying Savanna on their own cloud infrastructure.

Particularly noteworthy are the independent scaling of storage and compute, dedicated compute environments for OLTP and ALAP workloads, and high-speed data ingestion from new sources, according to Aslett.

Combined, they “provide greater flexibility to support varied and complex workloads,” he said.

While TigerGraph’s launch of Savanna aims to provide customers with a platform that provides the data discovery and performance needed to help build AI applications, market trends played a role in its development, according to Shrivastava.

Next steps

With Savanna now available, AI features prominently in TigerGraph’s product development plans, according to Shrivastava. In particular, like many data management vendors, TigerGraph plans to add and improve capabilities that enable customers to develop their own AI applications.

“We will continue to add new capabilities that enable customers to build AI applications with ease and agility,” he said. “These offerings will be consistent with our mission of delivering scalable and high-performing connected data analytics.”

Aslett, meanwhile, noted that graph databases are more of a complementary capability in AI development than a critical one.

Many developers are using vector databases to develop retrieval-augmented generation (RAG) pipelines that find data and feed AI models and applications. Like graph databases, vector databases enable similarity searches. In addition, by assigning numerical values to unstructured data, they enable developers to give structure to data such as text and images that otherwise are undiscoverable.

Graph databases such as Savanna can be an effective complement to vector databases in AI development by providing additional data via similarity searches, according to Aslett.

“Knowledge graphs … stored in graph databases can complement RAG processes by augmenting similarity search and GenAI models with additional information related to relationships that are not represented by features or attributes stored as vectors,” he said.

Eric Avidon is a senior news writer for Informa TechTarget and a journalist with more than 25 years of experience. He covers analytics and data management.

This is the hidden content, please

• Sign In
• or
• Sign Up

#TigerGraph #launches #Savanna #aid #development

This is the hidden content, please

• Sign In
• or
• Sign Up

This is the hidden content, please

• Sign In
• or
• Sign Up