Is Cloud-Fragmented Data Stalling Analytics Initiatives?

Studies and experts say the gap between the cloud’s promise and actual business results can be traced to siloed data across platforms.

By Stan Gibson

By Stan Gibson April 27, 2020

With many years of cloud computing under their belts, some experts are taking stock of how well the cloud is delivering on real-world expectations. According to a recent survey of 900 senior IT decision makers conducted by Vanson Bourne on behalf of data management company Cohesity, things could be better.

More than two-thirds (68%) of the IT pros surveyed said the cloud has fallen short in delivering anticipated benefits of simplified operations, increased agility, reduced costs and greater data insights. Of those respondents, 91% cited mass data fragmentation as the culprit.

Significantly, mass data fragmentation is standing in the way of perhaps the key transformative benefit of becoming a digital business: the ability to analyze structured and unstructured data for actionable business and customer insights. Indeed, respondents to the Vanson Bourne study identified the top benefit of removing data fragmentation as improved analytics/AI on massive data sets now stored across multiple clouds.

How Does Data Become Isolated?

Data fragmentation happens when different applications launch in different clouds – an Oracle ERP application here, a CRM application there, for example. If those applications need to use the same data, then multiple data copies are created and remain in the silo of each cloud-based application. Over time, multiple data copies become multiple data versions as changes are made to what was once the same data.

“If marketing were to launch an AWS server, there would be files in-house and on AWS,” said Brian Grime, CIO of Superior Credit Union in Lima, Ohio, envisioning how fragmentation would play out at his institution.


Things can get complicated quickly, added Roy Illsley, distinguished analyst, infrastructure solutions at Ovum Research.

“There can be duplicate copies, perhaps 15 different versions of a PowerPoint presentation across multiple locations,” said Illsley.

With data stored in different places, companies lose the single source of truth, said Holger Mueller, vice president and principal analyst at Constellation Research.

“If the HR system is in the cloud, but the financial system is on-premises, it will create problems. You can’t see the bigger picture,” he said.

[Related story: Get Ready for Infinite Computing]

He said a hiring manager would need to have up-to-date information about the number and compensation of current employees, as well as future financial projections in order to make a decision on whether to post new job openings. Failure to make prompt moves could put an organization at a disadvantage with faster-moving competitors.

“The holy grail is insights to action…if you are too slow to pull in HR data, 24 hours later, it might not be the same information,” said Mueller.

Zeus Kerravala, founder and principal analyst with ZK Research, summed it up this way in an article in Network World: “Fragmented data will lead to fragmented insights, which can lead to bad business decisions.”

Two Integration Options

According to experts, the solution comes down to either consolidating all data in a single database in the cloud or implementing an integration platform, which is a metadata layer or fabric that knows what the data is and where it is. Mueller said data consolidation has its advantages. With data in a single location, it will be up to date, and there will be no need to search for it across different locations, he pointed out. However, Illsley countered that for many companies, such consolidation is impractical.

“Data will be fragmented,” he said. “That is just the reality of our life. Good management of data is the answer.”

Almost nine in ten respondents (88%) to the Vanson Bourne survey were optimistic that the promise of the public cloud could be realized with a solution that would solve their mass data-fragmentation issues. Given that perception, where’s the best place to start in the search for greater cloud value?

“Don’t make things any worse. Whatever is there today, take control of it,” said Rahul Tripathi, vice-president and CTO for customer success at Nutanix.

He said it’s important to choose an integration platform specifically created to help organizations realize the promises of the cloud era. The Nutanix Acropolis OS, for example, runs across a range of hardware from multiple vendors and public cloud environments such as AWS.

Because up-to-date information is critical to informed decision-making, Acropolis OS is a significant improvement over earlier-generation integration platforms that become performance bottlenecks by forcing all data traffic to pass through them, said Tripathi.

“Acropolis OS implements Apache Kafka, a distributed streaming platform that does not create a chokepoint as businesses store and retrieve data,” Tripathi explained.

“Acropolis OS lets businesses host their data, whether structured or unstructured, with standard interfaces, effectively getting rid of silos and data fragmentation.”

He said this approach is particularly beneficial for big data analytics initiatives. He sees many organizations build data pipelines for big data analytics on top of the Nutanix platform to search across different types of data. The result is that data once again becomes a single version of the truth, a source of valuable insights into business data that can enable an organization to achieve new levels of competitiveness.

Stan Gibson is a contributing writer with 36 years of experience as a technology journalist.

© 2020 Nutanix, Inc. All rights reserved.  For additional legal information, please go here.