This is part 2 of our conversation with Forrester analyst Michele Goetz. Please click here to read the first post: Rethink Hybrid for the Data-Driven Enterprise.
After a recent Actian webinar featuring Forrester Research, John Bard, senior director of product marketing at Actian, asked Forrester principal analyst Michele Goetz more about next-generation data management solutions. Here is the second part of that conversation (see part one here):
John Bard, Actian: What are key business imperatives that are forcing a greater priority of speed of query processing for systems of insight?
Michele Goetz, Forrester: More and more businesses are becoming digital. Retailers are creating digital experiences in their brick-and-mortar stores. Oil and gas companies are placing thousands of sensors on wells to get information on production and equipment states in real time. And the mobile mind shift is driving more and more consumer and business engagement through mobile apps. Everything is in real time, delivered through a web of microservices, and increasingly sophisticated analytics are embedded in streams and processes. This places a significant demand on systems that have to hit high performance levels on massively orchestrated data services to get insight on demand, make decisions quickly, take action quickly, and achieve outcomes that meet business goals.
JB: How important is it for operational data and systems of insight to be tightly linked? What are some applications/use cases driving that integration?
MG: More and more, transactional systems have to operate on insight and not just as entry points to capture a transactional event. Analytics are running on streams of data and individual transactions such as purchases and business process events and transactions. These analytics are providing suggestions and instructions to inform pricing, offers, next best action, and security/fraud patterns, along with automate manual processes. Today’s modern data platform has to run analytic and operational workloads side by side to not only enable a process but also capitalize on opportunities and threats as they occur.
JB: How does an enterprise strike a balance between best-in-class solutions that often require integration versus all-in-one platforms that often force compromises?
MG: For each business process, customer engagement, automated process, and partner engagement, there are different service-level needs for data and analytics. Data and data services have to be more personalized to the tasks at hand and desired outcomes. Upstream in-development applications are designed with specific requirements for data, insights, and the cadence for when data and insight are needed. These requirements manifest within the data and application APIs that drive microservices and business services. A monolithic all-in-one platform creates rigidity as a purpose-built system that is inflexible to business changes. The cost to purchase and maintain is significant and has an impact on the ability to modernize, thus building up technical debt. Additionally, for every new capability, a new silo is built, further fragmenting data and inhibiting insight. Companies need to move toward a hybrid approach that takes into account the cloud, data variety, service levels, best-in-class technologies, and open source for innovation. Hybrid systems allow flexibility and adaptability to drive service-oriented data toward business value without the cost and delivery bottlenecks that one-size-fits-all systems create.
JB: What is the best design approach to accelerate development to achieve faster deployment to production and therefore business value?
MG: Start with what the solution is supporting and the service levels it requires. Have an understanding of how that fits into specific data architecture patterns: data science for advanced analytics and visualization, intelligent transactional data , or analytic and BI workspaces. These patterns guide the choices for database, integration, and cloud while also helping to establish governance that guides trusted sources, repeatable and reusable data APIs and services, and the management of security policies.
JB: What sort of new applications and services can be created from these new hybrid data architectures?
MG: Hybrid data management is about putting the right data services and systems to the task and outcome at hand. It provides more freedom to introduce modern data technologies to quickly take advantage of capabilities to scale, get to insights you couldn’t see because of lack of data access, and deliver data and insight in real time without the lag from nightly batch processing and reconciliation. Additionally, hybrid data management has better administrative layers to help manage the peaks and valleys across the ecosystem and avoid performance bottlenecks, as well as right cost data service levels between cloud and on-premises systems. Going hybrid means getting access to all the data to create customer 360s that take personalization to the next level. It allows analytics to mature toward machine learning, advanced visualizations, and AI by providing a better data infrastructure backbone. And apps and products become more intelligent as hybrid systems create engagement that is insightful and adaptive to the way the solutions are used.