How can organizations best prepare for the wealth of new data technologies coming their way?
There are various terms being bandied about that describe the new world data centers are entering—from the “third platform” to the “digital enterprise” to the “always-on” organization. Whatever the terminology, it’s clear there is a monumental shift underway. Business and IT leaders alike are rethinking their approaches to technology, rethinking their roles in managing this technology, and, ultimately, rethinking their businesses. The underlying technologies supporting this movement—social, mobile, data analytics, and cloud—are also causing IT leaders to rethink the way in which database systems are being developed and deployed. Mobile computing has shifted the balance of power between users and enterprises, while also increasingly becoming a primary vehicle for collaboration and communication. In addition, cloud now offers a wide range of services and applications for today’s data center.
The results of such rethinking may be profound, with IT and data leaders increasingly becoming engaged with the overall management of their organizations.
1. MOVE AWAY FROM ONE-SIZE-FITS-ALL AND BUILD-IT-TO-LAST THINKING.
Technology changes almost annually, with new solutions and approaches constantly emerging. A few years back, data warehousing emerged as the final step in data integration for analysis. Apache Hadoop then took center stage as the enterprise solution for capturing and maintaining large volumes and varieties of data in one place. Now, many organizations are exploring data lake approaches, in which data is maintained in its raw form for later processing as needed. The key isn’t to sign on and build around any of these technologies or approaches; rather, it is to maintain an architecture in which these new approaches can be plugged in.
These new platforms address the pain points seen in many data environments in recent years, especially in terms of the need for always-on availability, supporting volumes and varieties of big data, and rapid deployments to meet ever-changing requirements. No single type of environment, or architecture, can address all the needs of a 21st-century enterprise. A typical enterprise may have varying configurations of relational database management systems, Hadoop, in-memory systems, data warehouses, and analytics platforms. There are a lot of choices and options to get the right data to the enterprise that can be turned into actionable insights.
2. RETHINK THE WAY DATA FLOWS INTO THE ENTERPRISE.
For years, organizations pursued a very predictable path to data integration: raw data sources were identified; processed through extract, transform, and load environments; and archived in data warehouses. Now, data lakes take a whole different approach—data is brought in and stored as raw data with enterprises, for users and applications to handle as they want. In fact, data lakes may finally solve the problems that have vexed data sites for years. That is, all the big data streaming into organizations need not be transformed or tied to a schema. Ultimately, there may be more reliance on data maintained outside the firewall than inside.
This access to outside data—which may be coming from the Internet of Things, with its billions of connected devices, or from social media, or from cloud sites, potentially is too much for most internal corporate systems to bear. By bringing in data in its raw form and storing it in a centralized repository, it could be made available for analysis with whichever tools and platforms the users prefer. While Hadoop is often the platform of choice, it is only a piece of the picture. Raw data is brought in and stored in this centralized system known as the data lake. The data lake enables users and applications to quickly access the data they need, when needed. In addition, data lakes serve as a vehicle to support IoT initiatives. As IoT data streams in from multiple sensors, devices, and chips, the volume may overwhelm conventional data flows.
3. RETHINK EMPLOYEE HIRING, TRAINING, AND RETRAINING.
For years, data employees have tended to specialize in specific areas and on specific platforms—as Oracle DBAs, or DB2 developers, for example. Today, however, data employees require multiple skills. Developers may need to be administrators and vice versa. Many of today’s data professionals have the chops to be data scientists, of which there is projected to be a significant shortage. These new roles are still undefined and open to new ways to fill them.
4. MOVE AWAY FROM EVERYTHING-ON-DISK-AND-SERVER THINKING AND TO SOFTWARE-DEFINED ENTERPRISE THINKING.
A software-defined enterprise is built on three pillars: software-defined network, software-defined storage, and software-defined servers (otherwise known as server virtualization). The key is to think about data management as a capability that can be delivered across the enterprise in a virtual way versus physically building databases. As a result, the types of databases and disk arrays within the organization may no longer matter as much, as all would be equal partners within a software-defined setup.
5. RETHINK THE TYPE OF CLIENTS BEING REACHED. MOBILE DEVICES REPRESENT THE NEXT WAVE OF CLIENT DEVICES.
Increasingly, users of mobile devices can gain access to even the most powerful and sophisticated back-end applications —and data. Conversely, there is also a wealth of data coming out of these devices as well, which provides insights on user behavior and preferences. Experts calculate that the number of mobile devices is already poised to exceed the number of people on the planet. The challenge is to employ the data to enrich the customer experience, while boosting employee productivity and satisfaction.
The next generation of mobile apps —beyond Angry Birds and Trivia Crack —is focusing on enriching experiences for employees, while capturing and presenting metrics on everything from financial data to customer retention for enterprise decision makers. But the mobile experience is far different from the desktop PC experience. The question is, how to leverage the data coming in from mobile users and their engagements.
6. RETHINK HOW SOFTWARE AND DATABASES THEMSELVES ARE ACQUIRED.
Many of today’s cutting-edge data solutions are open source. Hadoop, for example, is an Apache project on which various vendors are adding components. Open source databases are also becoming as powerful and robust as their commercially available counterparts. This is part of a great commodization wave sweeping enterprises, in which organizations look to assemble solutions from components and software assembled and maintained by communities of interest. There are a variety of open source databases available, from relational to NoSQL, as well as open source frameworks and accompanying tools. This provides wider choices, and little or no licensing costs as data environments scale, as well as greater opportunities for developers to test and deploy for a range of functions along with mission-critical applications.
Business and IT leaders alike are rethinking their approaches to technology, rethinking their roles in managing this technology, and, ultimately, rethinking their businesses.