Accenture Insights Platform Now Offers Splice Machine
Accenture’s analytics-as-a-service solution adds Splice Machine’s intelligent application platform to enable customers to rapidly build and deploy predictive applications at scale
We are excited to announce the availability of the Splice Machine data platform as part of the Accenture Insights Platform (AIP).
AIP is an analytics-as-a-service solution designed to simplify analytics and support a data-native culture. It delivers actionable insights that unlock business value trapped in data and create new value for clients. Offered in a pay-as-you-use service model, the platform includes a suite of pre-built analytics apps and tools to support a multitude of industries and business functions.
The addition of Splice Machine to AIP not only enables customers to leverage data at any scale for operational and analytical workloads, but with built-in machine learning and a data science workbench, the Splice Machine data platform also enables customers to easily build and deploy mission-critical real-time AI applications with much less complexity. Splice Machine is a full-stack AI platform, requiring fewer infrastructure engineers to duct tape compute engines together and demanding much less data movement, so that analysis and machine learning are up-to-date and in the moment, resulting in smarter applications that make better decisions.
“Given the rapid adoption of AI, we want to provide our clients with a platform that can serve their needs today and in the future,” said John Matchette, Managing Director and Global Lead of the Accenture Insights Platform. “With Splice Machine’s scalability and performance on almost any SQL workload, the Accenture Insights Platform is now even more well equipped to deliver on these needs.”
Building sophisticated AI capabilities involves thinking through the three broad phases of turning data into actions based on predictions:
- Analyzing data
- Performing machine learning experiments
- Powering real-time, mission-critical applications that deploy ML models to make intelligent predictions in the moment
As a key component of the AIP stack, Splice Machine provides an integrated data platform to simplify and accelerate each phase.
The first phase entails how easily one can gather and analyze data to prepare for machine learning. Massive new data sources and streams, such as those from Internet of Things devices, are driving enterprises to collect more data than ever before. The AIP team chose Splice Machine for its horizontal scalability and full SQL functionality, allowing customers to ingest data with a high degree of parallelism in both a batch and streaming manner.
“We are excited to bring a next-generation AI solution to market with Accenture. As more and more businesses seek to digitally transform, deploying machine learning in the fabric of the enterprise workflow is critical for them to stay ahead of the competition,” said Monte Zweben, CEO and co-founder at Splice Machine. “Our completely integrated solution enables customers to not only to start realizing value from day one, but also to continually learn in real time – all backed by Splice Machine and Accenture’s proven expertise.”
To power the second phase of adding AI to the enterprise, Splice Machine’s architecture enables analytics at scale with Apache Spark embedded in the engine. It also supports ANSI SQL with a cost-based optimizer that can statistically determine data access via tables or indexes, as well as optimal join ordering and algorithms. With its native Spark DataSource, Splice Machine enables data engineers and data scientists to use the full power of Spark libraries such as Spark Streaming as well as Spark MLlib with minimal data serialization and movement. Data scientists can manipulate result sets as Spark DataFrames and durably update and store DataFrames in Splice Machine transactionally. Additionally, Splice Machine exposes SQL and Spark in pre-integrated Apache Zeppelin notebooks where data scientists can collaborate on the steps of machine learning experiments including feature engineering, model selection, parameter tuning, training and testing.
The system enables full consistency of highly-concurrent transactions unlike NoSQL systems and other scale-out systems. Splice Machine makes it easy to incorporate ML models directly in operational applications and enable frequent retraining of models since the analytical and operational engines are co-resident but use isolated computational resources.
Whether building powerful AI systems, operationalizing languishing data lakes with real-time applications, or just migrating under-performing workloads from existing platforms like Oracle or Teradata, AIP with Splice Machine helps address a wide spectrum of intelligent scenarios by providing a scalable, highly functional, and high-performance analytics-as-a-service platform.