Cloud computing has gone from being a promising technology to a reality that brings a unique set of challenges along with benefits. To fully leverage the disruptive potential of cloud without getting trapped in a web of integration complexity, CIOs and their IT organizations need to focus on what it means to rethink their business as a collection of services.
IT organizations now have a bewildering array of options at their disposal--private cloud, public cloud, hybrid cloud and, of course, on-premise installations. In most organizations, according to Deloitte Consulting's Mark White and Bill Briggs, adoption is no longer about cloud; it's about clouds.
"As cloud offerings added vertical business capability offerings to the horizontal IT capacity services, the adoption question changed from "if" to "when"--and the answer is frequently "now,"" White, principal and CTO of Deloitte Consulting, and Briggs, director and deputy CTO, write in their Tech Trends 2012: Elevate IT for Digital Business report, released last week. "Along the way, leading organizations moved from cautious exploration to the reality of multiple individual cloud offerings handling critical pieces of their business operations-and sourced from multiple public and private providers."
In other words, forget public cloud and private cloud. Hybrid cloud, composed of at least one public cloud and at least one private cloud, is where it's at. In fact, White and Briggs said businesses are increasingly willing to deploy multiple applications and infrastructure services on cloud platforms. Deloitte calls this the "hyper-hybrid cloud," in which multiple clouds must link back to the core and often to each other.
"In each instance, these offerings needed to be connected back to the core of the business, often through traditional data-driven on-premise integration solutions," they said. "Advance one step further and the organization is managing both exception and routine workflow across a growing range of disparate cloud offerings with point-to-point links to legacy systems and data. This shift from 'cloud' to 'clouds' provides new opportunities, but it also brings challenges beyond just integration-security, data integrity and reliability, and business rules management for business processes that depend on enterprise IT assets composed with one or more services."
While dealing with these challenges, IT shops need to keep up the on-premise side as well, because it's not likely to go anywhere soon, even in organizations that are eagerly consuming cloud services.
"Our perspective is that it's irrational to think that someone will go to work on Friday and all their work will be on-prem, and then they come into work on Monday and everything will be in the cloud," said Mike Ehrenberg, Microsoft Technical Fellow with Microsoft Business Solutions. "People are going to live with assets in both places for a long time."
Unless organizations have a plan in place, tying these disparate services and the on-premise component together can turn into an integration and orchestration nightmare, said White and Briggs.
"Integration, master data management and enterprise architecture have historically served as the linchpins in modern IT shops--a role that has only recently become more important with the adoption of multiple cloud solutions," they said. "As more functional business leaders independently subscribe to cloud offerings outside of the trappings of traditional IT, underlying business processes can become riddled with multiple cloud players that the organization itself must integrate and orchestrate. As a result, much of the "IT-free" value proposition can dissipate at the enterprise level."
Some organizations are turning to cloud services brokerages (CSBs) to deal with this problem. Gartner described just such an example in Case Study: Mohawk Fine Papers Uses a CSB to Ease Adoption of Cloud Computing, released in July 2011. Mohawk, the largest premium paper manufacturer in North America, has an IT staff of six, but needed to create an SOA-based backplane to support interoperability among all its internal and external applications, services and data.
"Mohawk considered a number of alternatives focused on cloud services integration because of its adoption of cloud computing, including providers of integration appliances and integration platform as a service," said Gartner Analyst Benoit Lheureux. "Citing concerns with the capital expense of integration products and its reluctance to hire more IT staff to do integration work, Mohawk decided to outsource all its integration work."
Mohawk chose to work with Liaison, an aggregator and orchestrator of cloud services that was able to manage the relationships and interdependencies required, leaving the back-end complexity transparent to Mohawk. Liaison provides Mohawk with on-premises integration, supply chain integration for its 300 customers and 100 suppliers and other external e-commerce partners and with intermediation of all its third-party cloud services providers.
"Mohawk has indicated that, by outsourcing integration as a competency, it was able to shift its strategic focus onto its core competency of managing business processes and master data (from its prior focus on acquiring integration technology and development work)," Lheureux wrote. "For example, when it needed reliable currency conversion to drive additional native-currency sales of its paper products, rather than on-premises or custom interfaces, it worked with Liaison to integrate and extend StrikeIron's foreign currency conversion service onto its SOA backplane. While Liaison focused on the technical implementation of doing that, Mohawk focused on successfully incorporating the external service into its various applications and processes to meet new business requirements."
Organizations that are ready to make the move to fully exploit cloud services need to adopt "services thinking," said John Seely Brown, independent co-chairman of Deloitte Center for the Edge.
"That's about defining operating models, business processes and technology components as services-within and beyond the enterprise," he said. "And it demands that companies develop a disciplined set of control points for an emerging services grid-control points with policies in place that flow up and down the stack. Unfortunately, getting a clear handle on the policy layer is often more easily said than done. Policies are often embedded across multiple layers of enterprise software, making them difficult to see, let alone orchestrate. It takes focus and resources to dig them out and make them fully explicit."
To successfully make this sort of transition, White and Briggs said it is essential that organizations first determine if cloud services models are even well-suited for their business problems and technical environments. They have a rubric based on the cloud definition of the National Institute of Standards and Technology (NIST) that they use to assess whether cloud is the correct fit for a business problem. The attributes are the following:
Predictable pricing: Subscriber incurs a charge only when consuming resources, based on subscription and usage, not on perpetual licensing or allocation
Ubiquitous network access: The service is available wherever and whenever the network is available-enterprise network for private cloud, Internet for public cloud
Resource pooling and location independence: Multi-tenant, with shared resources that the subscriber cannot explicitly partition or specify
Self-service:: Users can directly access the service, provisioning is on-demand, and services are readied in near-real-time
Elasticity of supply: Ability to scale up or down to meet resource demands, with seemingly limitless upper bounds
White and Briggs said that if all five of the attributes are relevant to a business problem, cloud is a potential fit.
"When preparing to move ahead, start by articulating the business goals, identifying key security risk and compliance considerations, and investigating potential cloud and traditional solutions," they said. "Speed to solution through rapid, low-risk implementation can be a benefit of cloud computing: get ready to think big, start small, fail fast and scale soon. The path to hyper-hybrid will likely lead to custom extensions of initial cloud services, increased interfaces to core back-office systems and data, and the pursuit of SaaS application capability cloud offerings."
Thor Olavsrud is a senior writer for CIO.com. Follow him @ThorOlavsrud.