Introducing IBM Cloud Private for Data: Your enterprise-grade AI cloud platform, working in your datacentre

One of the most significant benefits of building a cloud computing platform based on Docker Containers and Kubernetes is what you can do with it. I introduced in my initial paper [click here] the IBM Cloud Private Platform (ICP) which provides enterprise-grade capabilities for instant expansion in computing volumes, resilience, management, most importantly governance, and deep integration. We have taken advantage of these computing attributes to build a unique analytics platform we call IBM Cloud Private for Data (ICPd), which is IBM’s next-generation offering for Enterprises and is a specially engineered platform that serves up capabilities from IBM’s rich data and analytics portfolio, modernized for cloud nativity. ICPd is an engine for performing Machine Learning and Artificial Intelligence by working with potentially massive data sets. Additionally, ICPd gives you the ability to conduct extensive data science projects.

At the heart of the ICPd capabilities is a singular, highly governed data catalogue that will allow you to import data in massive volumes, from an extensive and disparate set of data sources both internal and external. The catalogue uses enterprise-grade data management and Machine Learning governance capabilities to curate the data thus at the same time ensuring that the data is of good quality and trustworthy. ICPd aims to solve the significant risk involved in feeding a machine learning algorithm faulty data that would lead to an artificial intelligence engine to reach wrong conclusions.

Imagine how catastrophic it would be for your business if you were trusting an artificial intelligence engine to infuse insight into your business processes and it made the wrong decisions.

A fact that makes ICPd a much needed product today, is that businesses and institutions are significantly more data-centric and data dependent. Because data is becoming the natural resource of the 21st century, it is imperative not just to create insight for its own sake, but to operationalize that data by infusing that insight into all your business processes. Adding insight into your processes is not a trivial problem to solve because to operationalize this desired outcome, you need to have analyze capabilities that allow Data Scientists and IT Operations professionals to work together seamlessly.

Therefore, an enterprise-grade system that will solve this type of problem needs to have robust Data Governance, Master Data Management, and Data Science capabilities built into the system. However, that is not enough. Such a system also needs to have very robust DevOps capabilities because the models that will be created by Data Scientists need to be operationalized into applications, some of which can be legacy applications, that run your business today. ICP for Data addresses this challenge by unifying all the key operations personas (Data Engineer, Data Scientists, Data Steward, Application Developer, Business Analyst & CXO, and so forth) in your organization, to get them to work as a unified, well-coordinated team.

As data sources and data volumes explode, ICP for Data comes with Data Virtualization capabilities already built into the product. With the Data Virtualization engine, you can query an unlimited number of data across many systems without having to copy and replicate data which reduces costs, and it also brings in that data into the catalogue to ensure the data is well organized. Consider that the bulk of all this data has real business value for periods as short as a few seconds to a few minutes, so speed is critical and the Data Virtualization engine solves that problem. This capability alone is a genuinely revolutionary advantage for companies that decide to compete by using data, it also simplifies analytics and makes it more up-to-date and accurate because you are querying the latest data at its source. This capability is only available on ICPd, because an engine of this type that presumably will be mining massive amounts of data creating large workloads, requires the computing robustness of the IBM Cloud Private Platform.

A final thought: Kubernetes serves as the foundation for ICP for Data and is designed to meet the expectations of enterprises to realize the benefits of both agility and cloud-like economics in the security of their controlled data centers. ICPd additionally because it works on Linux, can run on x86, Power Systems, Mainframes, or any system that supports this operating system. This flexibility of deployment should not be underestimated because for some organizations, you may require significant horsepower to successfully process large datasets on a real-time basis to capture substantial market opportunities, and you will not be disappointed given how flexible IBM Cloud Private for Data is as it pertains to all the platforms on which it works.

Thank you.

This article was written in collaboration with Christopher Smallwood and David Ma, Business Analytics Technical Specialists, IBM Canada.