While public cloud solutions have ?long been the preferred option for startups, existing organizations are also recognizing the advantages of such services. Most startups find it easy to ramp up by consuming public cloud solutions at first, and slowly migrating data back on-premise for cost optimization, Yadav said. However, more mature businesses always had data on-premise, and they migrated some of that data to the public cloud in order to reduce the number of resources needed to manage their infrastructure.
“Public cloud and ‘as-a-service’ solutions have transformed the expectations such that high-availability, simplicity, and security are expected to be built-in for consumer as well as enterprise products. It’s not enough to just fulfill the customer requirements anymore, delighting the customer is the key,” he noted
For the year ahead, expect to see “a huge push in data management vendors providing deeper public cloud integrations and removing the complexity out of hybrid cloud data management and analytics features,” Yadav said.
Itamar Ankorion, CMO at Attunity, predicted greater use of major cloud infrastructure and platform providers—such as Amazon S3, Azure Data Lake Store (ADLS) and ADLS Gen2, and Google Cloud Storage—for cloud-based data storage, especially as organizations move to analytics-driven strategies. “These converged platforms host preferred analytics systems such as the Snowflake and Amazon Redshift data warehouses, and Amazon Athena,” Ankorion pointed out. “These systems incorporate familiar SQL structures.”
The traditional cloud benefits apply here, Ankorion continued. “Enterprises gain economic storage, processing and management, elastic resource consumption, and the shift of CAPEX to OPEX. They also can reuse storage for new analytics workloads. All this is changing how organizations manage data because they can more cost-effectively address more advanced analytics use cases and thereby realize a greater return on their data investments.”
However, there will be a some challenges in achieving these cloud storage benefits, Ankorion cautioned. “Transferring high data volumes across wide area networks can result in high latency and consume costly bandwidth,” he explained. “Enterprises also need to avoid lock-in and maintain full data mobility, moving workloads between clouds and across hybrid architectures based on lessons learned and changing requirements.”
IT’S HYBRID TIME
Many enterprises have on-premise data assets they want to keep within their four walls, at least for now. With the rising popularity of public cloud, “2019 will be the year of multi-cloud and hybrid cloud,” said Arun Murthy, CPO and co-founder of Hortonworks, which recently announced a merger with Cloudera. “Cloud providers will, more aggressively, differentiate among each other in specific areas—such as operational readiness at Amazon, enterprise integrations for Microsoft, and AI and machine learning at Google,” he predicted. This array of capabilities from across vendors will give rise to multi-cloud strategies being put into motion, bound by “common security, governance, and data or workload management strategies. We will also start to see enterprises move some always-on workloads from the cloud back on-prem for a hybrid model to optimize economics.”
STRENGTHENING THE DATA FOUNDATION
As data becomes the lifeblood of most business models and strategies, expect to see a boom in the use of algorithms throughout the enterprise for a variety of purposes. “This has the highest opportunity for business payoff, especially for augmenting human intelligence,” said Suketu Gandhi, partner in the digital transformation practice at A.T. Kearney. He predicted that as a result, “data and organizational intelligence will be discussed at the board level, and CEOs will be talking about it regularly as well.” Gandhi cautioned, however, that data managers need to get ready for the push to an algorithm-driven enterprise. “It shines a light on the state of data— both internal and external—and the ability of an enterprise to use it. There is a lot of work to get ready for the fundamental reality of structured versus unstructured, internal versus external, real time versus batch,” he added.