In addition, enterprises must take care to ensure that data is maintained securely across the diverse platforms and environments seen with hybrid arrangements. “Hybrid data environments present the risk of copies of data proliferating across on-premise, private, and public cloud instances,” said Anupam Singh, general manager of analytics at Cloudera. “Companies need to invest in compliance software that helps them track the data in every copy, transformation, and usage pattern.”
Managing costs and resources will be another challenge. “As resource procurement moves away from central teams, enterprises will find it hard to do capacity planning between fixed and transient resources,” Singh pointed out. “Companies need to invest in capacity planning tools before they are unpleasantly surprised by exploding expenditure in the cloud.”
Awasthi feels many of the issues with hybrid cloud computing are being resolved from a technical standpoint. “The main challenges with standing up and managing a stable and secure hybrid cloud environment are related to connectivity and integration,” he explained. “Most of the platform as a service cloud vendors offer platform services to make connectivity easy and secure. Also, integration is becoming less of a challenge because of the general push toward using open APIs that facilitate the transfer of information between cloud and on-premises systems.”
The management and organizational aspects of moving to hybrid cloud are still challenging, however. Enterprises moving to the cloud “must apply appropriate diligence to understand the impacts migration will have,” said Cherrington. “First is deciding which applications are best suited for a shift to the cloud. Ceding control of the infrastructure for the primary revenue generation capabilities of the enterprise must be approached with due care. Moreover, special attention must be applied to the ongoing management of applications shifted to the cloud as they are inevitably part of larger business processes that span silos of technology. Ensuring that the shift to the cloud does not introduce new latencies or weaken the linkages for mission-essential business processes is paramount.”
At the same time, hybrid cloud arrangements need the flexibility and adaptability to today’s business environments. “The nature of cloud requires frequent and fast changes to its technology and applications, which means organizations need to be working with a cloud service that is designed to move at the same pace—and to bring them along for the ride,” advised Steve Wood, chief product officer for Dell Boomi. “Hybrid data environments also force organizations to stay just as agile internally, which can be a disruptive challenge in the initial development stage until it develops into a streamlined process that allows IT investments to function at their optimal level.”
Still, data managers are likely to find the challenges within hybrid cloud environments to be familiar from issues they wrestled with earlier in their careers. “From a data management standpoint, the biggest obstacle is the same with hybrid as it is with widely distributed on-premise database deployments—how do you support an active-everywhere model in a practical, efficient, and cost-effective manner?” said Schumacher. “Legacy RDBMSs and most NoSQL platforms are either master-slave or multi-master in nature, which can make hybrid success tough if not impossible for numerous use cases. Getting around that difficulty involves ensuring the data platform chosen is multi-home, masterless, active-everywhere capable, and that it’s baked into the architecture from the ground up.”
Plan Ahead
The bottom line is that a hybrid database cloud initiative requires some planning. “An organization taking a more intentional approach makes a conscious effort to distribute workloads across clouds—both public and private—and between various public clouds,” said Rugg. “This is intended both to match the workload to the cloud that supports it best, as well as to provide some level of arbitrage across public cloud vendors.”