Technologies come and go, but the problems they hope to solve have a knack for endurance.
Back in the 1980s, information systems managers struggled with "islands of automation," presiding over a kludge of mainframes, departmental minicomputers, PCs and local area networks. Isolated system silos expanded due to the lack of cross-system linkages.
Passing data from one system to another typically involved manual data entry. Effectively harnessing the breadth of technology to address complex business problems was nearly impossible. Systems integration emerged from the chaos of disparate technology assets.
Three decades later, cloud computing platforms, -- SaaS, IaaS and PaaS -- have proliferated as organizations adopt multi-cloud environments. Cloud integration services, which aim to provide greater cohesion among cloud-based IT resources, now follows a path similar to systems integration. Cloud consultancies and other channel partners aim to support customers with this phase of cloud evolution.
A distant mirror
Eran Gil, CEO at AllCloud, a cloud professional services company based in Denver, noted the similarities between the dawn of systems integration and arrival of cloud integration.
"I think it mimics very well what has happened in the past in terms of disparate systems," Gil said.
Though we are in the early stages, we have entered a phase of multiple clouds, according to one expert. "When you overlay SaaS and IaaS hyperscalers into this, most clients already have multiple cloud providers," said Nicholas Merizzi, principal of Deloitte Consulting.
He said Deloitte works with customers simultaneously scaling SaaS and IaaS capabilities. In IaaS, "most are still focused on getting it right with a single primary cloud provider, but, in parallel, establishing strategies to operate and integrate multiple clouds."
The current trend differs in the rate of technology adoption and, of course, the integration venue. As for the former, organizations can acquire cloud services far faster than they could fill a data center with physical gear. The potential points of integration will only multiply over the next few years, given the accelerating cloud shift.
"We are definitely in the mainstream adoption of cloud, but there is still a long way to go," Gil noted.
The integration points, meanwhile, shift from in-house operations to external clouds. Much of the integration work will occur at the platform layer, sandwiched between hyperscalers' cloud infrastructure and SaaS applications, Gil said.
AWS and Salesforce, for example, are beginning to meet in the middle, as one builds upon its platform from the infrastructure layer and the other builds from SaaS, according to Gil. The companies earlier this year introduced a plan for tighter AWS-Salesforce integration.
Such cloud linkages will "solve problems that are still not solved by either one of those separately," Gil said.
Cloud integration services obstacles
Integration and management across clouds, however, remain difficult and complex to achieve, Merizzi said. Cloud rivalries are one issue. While some cloud providers pursue better integration, the competitive landscape generally tends to deter collaboration.
"Cloud providers seek to compete and gain market share rather than to cooperate and integrate," Merizzi explained. "This is not new in the world of IT, but rather is now surfacing across an ecosystem of new providers, leading CIOs to pick who their strategic partners will be."
Disparate standards, which flow from competing vendors, provide another integration obstacle.
"Each cloud provider follows its own standards and best practices for managing their technology stack -- such as application, data, infrastructure and security," Merizzi said. Cloud-native design, while desirable, can result in applications harboring built-in dependencies on the underlying cloud vendor, he added. Those dependencies further limit interoperability and portability across clouds.
In addition, distributed organizations that replicate their traditional IT environments in the cloud might end up duplicating disjointed systems, said Paul Wilkinson, executive vice president at 1901 Group, a Reston, Va., MSP and wholly owned subsidiary of Leidos. The company's federal customers, however, put a lot of thought into migration, creating cloud architectures that promote integration, he noted.
Eran GilCEO, AllCloud
Working toward integration
Merizzi said his clients tap open-standards-based technologies to achieve better integration across multiple clouds. The Kubernetes platform, for example, manages containers across public clouds. Other major standards include Docker, Kafka and Spring, Merizzi said, noting that list is not exhaustive.
"Most hyperscalers have embraced those standards and it enables some degree of portability," Merizzi said.
Deloitte also advises customers to design for living across multiple clouds. Merizzi cited master data management, application deployment, identity-based security, and cross-platform governance as the key underlying services for enabling multi-cloud environments.
The market's evolution, meanwhile, will improve integration. In the coming years, common underlying cloud services will commoditize through open APIs, Merizzi said. "This should drive cloud hyperscalers to embrace openness to integrate services across other clouds [and] achieve interoperability," he added.
But it's not just cloud providers experiencing the need for integration. Cloud consultancies also feel the effects of heightened customer demand. Gil cited increased collaboration across the company's three practices: AWS, Salesforce, and data and analytics. "It's blurring the lines," he said of cloud integration.
But connecting clouds will help customers address their most pressing business concerns. In sales, for instance, organizations want to know which opportunity surfacing from the marketing department will quickly close, Gil said. But without better cloud integration, those businesses may lack the data volume and machine learning capabilities to answer such questions.
"The integration piece of these different platforms, be it IaaS or SaaS, is really the secret sauce to bring significant insights to the organizations we work with," Gil said. "It's not just one or the other."