The ability to quickly act on information to solve problems, make decisions, or extend offers has long been the goal of many businesses. However, it was not until recently that technologies were available to address the speed and scalability requirements of real-time analytics.
At Data Summit Connect, a free 3-day series of data-focused webinars, a session titled "Enabling Real-Time Analytics," covered real-world strategies for leveraging data when and where it is needed.
As part of this session, Navinder Pal Singh Brar, senior software engineer, Walmart Labs, discussed how the retailer's customer backbone team developed a platform that enables extraction, transformation, and storage of customer data to be served to other teams in his presentation, titled "Building a Real-Time Multi-Tenant Data Processing and Model Inferencing Platform."
To watch a video of Navinder Pal Singh Brar's presentation, go here.
Each week, 275 million people shop at Walmart, generating interaction and transaction data. Taking in data from various channels and maintaining a uniform identity of each customer is necessary.
The company is utilizing Kafka Streams. The Walmart team is relying on Kafka Streams because of its simplicity, it is a library and not a framework, has an embedded database, enables interactive queries, and is highly scalable, among other attributes.
Following the motto of: "Depth, Freshness, & Reach," the goal was to have a customer data platform to process events, derive inferences, and serve knowledge in real time, not just raw data, that was reliable, highly available, and scalable, and offer high throughput and low latency to serve real time use cases, and have a pluggable design to onboard new models easily. The company's distributed streams processing platform built on top of Kafka Streams is capable of data ingestion, data transformation, feature extraction, model inferencing/sorting and post processing.
In addition, Vikas Mathur, vice president, strategy, and Avalanche general manager, Actian, covered the evolution of data warehousing over the years going from appliances, to appliances in the cloud and elastic clouds, and why a new approach is needed now in a presentation, titled "New Thinking in Adopting a Hybrid Cloud Data Warehouse for Your Organization."
To watch the video of Vikas Mathur's presentation, go here.
Actian's Avalanche is a hybrid cloud data warehouse that offers performance at scale, hybrid deployment spanning public cloud, virtual private cloud, and on-prem, offers out-of-the-box connectivity to SaaS apps, offers low TCO with cloud economics, and is offered as a fully managed service.
As an example of a real-time use case, Mathur cited AA, a U.K. auto insurance and services provider, that needed a fast and automated pricing and counter-fraud process that brought together policy data, claims data, and quote data for analysis. Using Actian, AA has improved the speed and efficiency of its counter fraud department and improved its understanding of fraudulent activities.
According to Mathur, a recent survey of IT decision makers shows that, despite the greater adoption of cloud platforms, much enterprise data still resides on-prem, making a universal approach to data access highly desirable. Avalanche Federated Query, coming later this year, enables a global view of business data for querying without expensive data movement to support a comprehensive view and adherence to data governance policies. "This is an example of being truly hybrid and going one step further than what everybody else is doing," said Mathur.
Webcast replays of Data Summit Connect presentations are available on the DBTA website at www.dbta.com/DBTA-Downloads/WhitePapers.