Reimagine data processing applications for analytics using Xcalar’s cloud-scale relational compute platform

“Xcalar’s mission is to provide a platform for developing & operationalizing complex business logic with AI and ML to analyze and manage the world’s information with simplicity, speed and scale”

- Vikram Joshi, Xcalar CEO and Founder

The future of analytics is software defined. Shed the rigid confines of contemporary data warehouses and define both schema (relational/ declarative SQL) as well as business logic (compute/imperative Python) through our Relational+Compute approach. Data Warehouse as Code (DW as Code) enables software to build malleable data warehouses that can adapt quickly to suit new requirements

Businesses need a robust platform to write and maintain distributed data processing applications that can work with data directly off the data lake

In current application development environments, the following pain points are typical:

High personnel costs: specialized, and therefore scarce, programmers and consultants are needed

Limited developer productivity: without a classic development IDE for distributed applications

Develop > test > operationalize cycles are slow - tending towards months to develop, weeks to test, and then months to operationalize

Once models are developed and tested within development sandboxes, developers often have to wait for IT to step in and tweak the models to optimize them for running on full production data

Managing change cycles at the source is difficult once applications are in production

Data Latency: data seen by user dashboards is often not current enough because of elaborate ETL processing cycles

Less than enterprise-grade production readiness: Managing robustness, reliability, failures, and other issues such as detecting data skew is hard

With traditional data warehouse silos, it is getting harder for organizations to keep up with data exploding to the massive scale of data lakes

Many organizations find it difficult to migrate away from specialized hardware-based data warehousing silos. They would much rather have the flexibility to move to any public or private cloud, or a hybrid thereof, and work directly off the data lake, while still being able to use familiar tools, with relational programming, including natively invoking SQL, calling embedded SQL, using loops and conditionals, and handling transactions and durability

To operationalize their ML models, businesses find that they are having to provision for enormous compute resources and inordinate amounts of time to process petabytes of data

Once ML algorithms are modeled with open source libraries like TensorFlow, Spark ML and H2O, it is hard to operationalize them. Without a platform that targets resource cost reduction, organizations must carry high TCO with cloud cost models that are based on CPU time and number of I/Os

Xcalar provides the most scalable, cost-efficient, enterprise-grade platform in the industry for all stages of an enterprise’s data pipeline and line-of-business applications.

The Xcalar Data Platform Ecosystem

The Xcalar Ecosystem


Xcalar’s next-gen platform for scale-out data processing applications and operationalizing ML

Allows for ML models built with familiar open source environments, like TensorFlow, Spark ML and H2O, to be quickly operationalized at cloud-scale - billions of classifications in seconds

Has a sophisticated visual studio and IDE that does not require retraining for business analysts, DBAs and programmers, thereby saving training costs and accelerating develop > test > operationalize cycles

Leverages existing investments in infrastructure and in open source standards, languages and tools, including Hadoop, Spark, Tensorflow, etc.

Enables large groups of business analysts to concurrently use familiar BI tools and reporting applications to access historical and constantly updating, transactionally consistent data in real-time


The Xcalar Product Architecture


The Xcalar Product Architecture

Key Use Cases

Developing and operationalizing complex business logic and ML algorithms at cloud-scale

Interactively work with petabytes of data; billions of classifications in seconds

Virtual data warehousing and ad hoc analytics

Migrate your traditional DWs to an open scale-out architecture

Data transformation and quality at petabyte scale

Work with trillion rows; process petabyte scale batch data in seconds or minutes

Live data for BI and reporting

Unlock your data lake for your analysts and data scientists