Automation Takes on the Heavy Lifting of Data Management

<< back Page 2 of 3 next >>

Mathias Golombek, CTO of Exasol, sees automation being employed for “event-stream processing databases that have to instantly react in case of certain events.” Golombek cited the example of an alarm monitoring system, where sensors send data to a central entity that has to react as quickly as possible in case of a fire hazard, or a stock exchange news stream analysis system that can trigger automated buys/sells of certain stocks.

The most common forms of data automation are being seen in databases that are provisioned automatically, said Monte Zweben, CEO of Splice Machine. “It used to take days to months to provision and install a petabyte-scale distributed database. Now through container orchestration in the cloud, you can provision and install in minutes.” In addition, automation is leading to databases that now self-heal, he added. “Distributed databases systems now can shard or distribute petabyte-scale data across multiple machines and deliver high availability by automatically monitoring all the deployed containers, perform health checks, kill unhealthy containers, and auto-scale new ones—all with zero-human operator involvement.”

Another area ripe for data automation is in “analyzing large amounts of data and getting automatic insights out of it,” Golombek continued. “There are various applications, from AI algorithms using deep learning methods to win games against human beings to picture recognition software, or applications that automatically survey and optimize a company’s production chain.”


The technical problems that can be solved by automation are self-evident, then. But what business problems are suitable for data automation? It boils down to making everyone across the enterprise more productive. “People should realize that they are better off with a system that doesn’t require a complex task than a system that automates a complex task,” said Joe Pasqua, EVP of products at MarkLogic. “Today, the most commonly automated database tasks are in some sense the most prosaic, like deployments, upgrades, backups, and change testing. Those are all important, but are primarily operational rather than providing direct business value.”

Schumacher pointed to staffing and skills issues as a prime driver for aiding the business at large, given a “lack of operations headcount or lack of skills and experience at knowing what’s needed to maintain sophisticated database deployments.” In addition, automating database tasks “increases the productivity of the operations staff and allows them to focus on other things,” he noted.

For a business that intends to run on data, this becomes key. “Numerous business decisions have to be taken every moment, 24 hours a day, and database automation helps to solve this problem,” said Golombek. “By automating decisions with intelligent algorithms and analyzing comprehensive data across the organization, businesses can reduce the number of mistakes, increase the complexity of decisions made, and minimize the reaction time.”

Data management helps people make better decisions by using better data, said Jake Freivald, VP of marketing at Information Builders. “This is especially important for AI solutions, where ‘black box’ algorithms don’t allow businesspeople to see the process that gets them to an answer. Small errors in an AI situation can make business users doubt the results they get.” Moreover, Freivald said, businesspeople really “own” the data. “They know how critical specific pieces of data are, and what the consequences are for having bad data. They’re responsible for fixing data when it’s broken. IT can only help so much with those things, and typically only in generic ways, before the business has to get involved. Without automation, data management involves businesspeople paying close attention to many different data points and looking for anomalies. If we can automate the anomaly-identification process, we can help businesspeople spend less time looking for bad data and more time fixing it.”

Chadd Kenney, VP of product marketing and solutions engineering at Pure Storage, sees reduced application time-to-market as an important business driver for data automation. Whether DBaaS, on-premise, or in the cloud, automation “gives you the ability to rapidly develop, test, and deploy new applications at a fraction of the time it would take with legacy systems,” he said. “Enterprises can take new applications to market much faster, gain an edge over competitors, and service customers better. In addition, “storage administrators will no longer need to spend time and effort on provisioning storage resources for databases. DBAs will no longer have to wait for database resources to be allocated and provisioned. This frees up time for both teams to focus on more strategic, higher-value activities.”


Yet, there are areas of the data center and the business that may be out of the reach of automation for now, for a myriad of reasons.

“Much of database automation is about storing data, and maintaining data, but less about exacting significant business insights out of the data stored in the database,” said Ryohei Fujimaki, CEO of dotData. “As we all know, data is only as valuable as the business insights it produces. From that perspective, the current database automation technology needs improvement in this area.”

The majority of business tasks are still out of reach for automation, Golombek agreed. This is partially because there is not enough data collected, and automation algorithms—such as machine learning—are simply working to solve way too many problems, Golombek noted. “Databases can address only tasks with relatively simple rules and limited complexity. Compare that with robots that can take over many manual tasks within a car factory, but for the tasks that require soft skills, emotional intelligence, and a far more complex way of communication, robots are still not a fit.”

<< back Page 2 of 3 next >>