Bringing Order to Database Ecosystems

Order is the central function of a database, which stores information in ways that make it easy to find and use. But the database landscape is anything but orderly. Indeed, it has evolved into a jungle of formats, features, engines and applications.

Tobias Ternström has built a career on mastering the nuances of intricate database ecosystems. He’s the vice president of engineering and product at Nutanix, whose software helps developers and enterprises build digital infrastructures to support their computing environments — and, naturally, their databases.

Nutanix helped pioneer hyperconverged infrastructure, which uses software to virtualize compute, storage and networking functions needed to run a data center. Today, the company puts a lot of energy into helping its customers implement hybrid multicloud environments.

Combining on-premises data centers with public clouds from a raft of providers is difficult enough. Toss databases into the mix and things get really tricky. This is where Ternström steps in to help. In an interview with The Forecast, Ternström said the companies he works with have straightforward objectives.

“They want to reduce the risk, effort and cost of operating 100s to 1,000s of databases,” Ternström said. 

Database complexity becomes a formidable challenge, so how should today’s technology leaders ensure they’re getting the most return on their database investments? While there may be no silver bullet approach, Ternström said there’s always a way forward.

“It’s critical to step back, focus on the right strategy and build from there then optimize,” he said.

A quick review of his career and thoughts on emerging database trends reveals the kinds of database-centered issues technology leaders are grappling with.

One Man’s Life Among Databases

Ternström grew up in Stockholm, Sweden, where his techie dad introduced him to computers. Bitten by the PC bug since boyhood, he started his career in the 1990s dot-com boom, coding with SQL databases for a tech startup. The startup was self-funded, so Ternström taught SQL database classes to pay for rent and groceries.

“I used to code in the evening and teach SQL Server during the daytime,” he recalled. 

In the ensuing decades, he’s delved deeper and deeper into the nuances of database engines, formats and architectures. In a world with growing complexity, he’s maintained a simple idea of a database’s purpose.

“Databases have been around for thousands of years,” he said. 

He explained that any book with an index is a database because it stores information and gives people a practical way to find specific data bits.

The earliest electronic databases performed basic storage and retrieval duties. Database complexity grew in tandem with rising computing power. The internet and cloud technologies shoved everything into hyperdrive.

Why Today’s Database Landscape is so Elaborate  

As computers do more things, people demand more from databases. 

“It’d be great to have one database,” Ternström said, but that’s impractical because every application uses data in different ways. Moreover, keeping everything in a central database slows development, hampers scale, and increases risk.

Over the years, database engines like Oracle Database and SQL Server gave developers and enterprises many more options. So, IT leaders thought about trying to standardize operations on a single database engine. 

“But that also turned out to be tough because as you're building applications, different ones have different requirements,” Ternström said. 

Issues like cost, scale and features drive developers to the database engines that perform best with their applications.

He painted a picture of today’s world filled with microservices and cloud-based, as-a-service apps. 

“You'll find that a single microservice may use one database while another microservice in the same app is using a different database,” Ternström said.

Nobody has one database (or even a dozen) anymore. 

“They have hundreds, to thousands, to tens of thousands of databases,” he added.

How Automation Drives the Evolution of Databases

Database titans like Oracle will keep building proprietary databases for enterprises. Open-source databases will keep growing in use. And all these will have to be designed, secured, backed up and fine-tuned for peak performance.

These responsibilities pose mountains of difficulty when databases number in the thousands. Hence, many organizations turn to database automation, using computer science to replace manual tasks like software patching and operating system upgrades.

Ternström said automation frees up database administrators to focus on high-level business challenges and avoid getting bogged down in repetitive tasks. Automation also can be a lifesaver for things like backup and disaster recovery. 

Those are the upsides. But Ternström and his fellow DB aces have a lot to figure out. Every DB automation must work across multiple DB engines, operating systems and IT infrastructures. 

“Automation is not black or white,” he said, because of potential downstream effects that can break things later. 

“You want to automate your least critical databases first,” he added. “And generally, the most critical databases would be the last that you automate.”

Some tasks, like figuring out how much data loss is tolerable in a real-time disaster, require a human to make the right call, Ternström said. 

But overall, he recommends embracing automation where there’s the best balance of risk and reward.

Database as a Service: The Next Phase of DB Automation

Developers no longer need to master all the intricacies of databases, including backup, security, high availability, etc. They can automate those tasks by partnering with a database-as-a-service (DBaaS) provider whose APIs let them plug fully functioning DBs into their apps.

Developers and IT managers want DBaaS to work across multicloud environments. So Ternström explained the four primary DBaaS options for hybrid multicloud operators:

• Dedicated DBaaS providers are often single database engine solutions. If the IT team needs to support different engines, they now have to deal with different vendor APIs and consoles.  
• Public cloud providers that include DBaaS in their technology platforms. These DBaaS are specific to the cloud provider, meaning you’ll need to use a different DBaaS for each public cloud. Nor do they work with databases running in on-premises datacenters.
• Small vendors that specialize in multicloud DBaaS. These services may limit their options to the public cloud, which rules out companies needing on-prem capability. Most only support open source databases, neglecting commercial databases, which still make up the bulk of deployments today.
• A full-featured hybrid multicloud DBaaS, which can manage all the major database engines across on-premises datacenters, colocation or edge locations, and each of the major public clouds from a central software platform.  

The fourth option is the rarest. 

“DBaaS is not where customers need it to be,” Ternström lamented because few software vendors provide true hybrid multicloud DBaaS.

Thriving with DBaaS Comes Down to Control

It turns out there’s a Catch-22 when it comes to managing databases. Admins can have control or automation, but not both, Ternström said. 

By control, Ternström meant the ability to access and customize database servers, such as selecting the operating system and database version you want to use.

“With a traditional infrastructure-as-a-service or three-tier approach, you get tons of control,” he said. “You can install whatever you want. But, any automation you get, you have to build yourself.”

Most DBaaS, by contrast, deliver automation but restrict users' control, he said. IT teams can get stuck with the operating systems, database version and extensions that the DBaaS vendor chooses to support - or not.

“You don't have to worry about it,” he added. “But if you need any of those things I mentioned around control, you can't have it.”

This is where Ternström turns to Nutanix Database Service, which can manage the most popular database engines (Oracle Database, MySQL, SQL Server and MongoDB), in hybrid multicloud environments. 

“Nutanix Database Service gives you full control with the highest level of automation across the most popular databases, wherever they are running.”

Ternström has carved out a niche in the rapidly evolving database ecosystem. 

“I see things in terms of databases,” he said. “If I’m at the airport check-in and something goes wrong, my mind fixates on what could be going wrong with the airline database.”

From books to blogs to enterprise IT operations, he’s fascinated by the concept of organizing information and making it useful. 

“There's a database for everything, cataloging absolutely everything,” Ternström said. 

It’s proof enough that he’s on the right path, one that will help bring more order to a world that’s increasingly going digital.

Tom Mangan is a contributing writer. He is a veteran B2B technology writer and editor, specializing in cloud computing and digital transformation. Contact him on his website or LinkedIn.

© 2022 Nutanix, Inc. All rights reserved. For additional legal information, please go here.