CloudBees sponsored this podcast.
With CI/CD becoming more and more the defining competitive edge in the marketplace, the big question is how, exactly, does one accomplish continuous delivery? On this episode of The New Stack Makers podcast, we sat down with Jack Norris, to talk about a big piece of that puzzle. He’s senior vice president at MapR, a company that provides a data fabric layer to the tech stack necessary for continuous delivery.
The concept of a data fabric, which allows data to be accessed seamlessly across all locations from databases to the edge, is a fundamentally different architecture, Norris said. Back in 2008, when MapR was created, it represented a paradigm shift in thinking about data. Traditionally, the application dictated how the data must be organized and you ended up separate application stacks and data silos.
In today’s world of continuous delivery and speed of application use, data is a big pressure point, said Norris. A data fabric is an alternative to these separate silos. The fabric can contain multiple types of data, support different types of workloads, stretch across locations globally, and provide easy access to data on premise, he said.
The lifecycle of data is also really important, Norris said, and has changed along with the explosion of technology over the last ten years.
Persistence is more than just storing the data in a database, but having that data available, sometimes immediately available, for analysis.
For example, a fraud detection application needs data from several sources to determine if the card being used is fraudulent. Once the card is swiped, the chip is read, the customer data is returned and any anomalies are detected before the purchase is approved.
From a data standpoint, the event comes in, you have to recognize the customer, gather historical data, look through a series of database transactions for anomalies for that specific customer in that specific location, all before the purchase is approved.
Today’s time frames for applications are super fast: as the web page is loading, as the credit card is being swiped. The persistence at the fabric layer is to support that speed.
The latency and need for speed are driving the need for the data fabric functionality, Norris explained. The MapR database not only does a key-value store, but it handles that natively, meaning that the data file doesn’t have to be flattened to be transferred, a process that reduced data granularity in exchange for the speed of moving enormous files.
In this Edition:
1:32: Data fabrics, what they are, and how they help the whole CI/CD process.
4:54: Defining data persistence: What is it and why does it matter?
7:52: Data availability and usage over time.
10:25: Exploring MapR’s origins.
15:58: What’s the hardware and the software that you use at MapR in order to make the data fabric possible?
18:52: How MapR is helping C3 IoT use data fabrics to support AI.
Feature image via Pixabay,