Maximizing Cloud’s silver lining - An integration love story

Hurricane Sandy demonstrated the value of Cloud as excellent for disaster recovery, but there is so much more to Cloud that isn’t always clear. I had the great pleasure of hearing Gartner’s Massimo Pezzini present at the Gartner ITxpo in Barcelona, Spain today.

Why Cloud integration?

Pezzini pointed out that organizations today play in multiple virtual enterprises that include supply chains and networks. He reminded us that SaaS systems already hold some of most enterprises’ data. Chunks of business process are already in the Cloud and lastly, that your application portfolio will include on-premises and cloud apps. Talent acquisition, CRM, the list is growing…you can’t avoid Cloud.

Beyond these reasons, organizations must share information with business partners and participate in their business processes. This trend means data will be fragmented and everyone’s end-to-end work will move on and off premises as business is executed. It’s here.

To be effective and efficient, you need to be ahead of the Cloud curve.

Benefits

Pezzini described two scenarios that dominate the Cloud environment. Regardless of the topology, the benefits are the same:

Cloud to Cloud: All the end points of today’s computing are cloud services. They can be a combination of SaaS applications, Web API’s (with your partners) and others (cloud info services).

Cloud to ground: End points are cloud and on-premises and are a combination of Cloud services and on-premises custom or packaged apps.

What changes, what stays the same?

There are things that stay the same in a Cloud environment, like data consistency/synching, multi-step processes that span apps and composite applications. Messaging layers, data transformation, and adapters all are just as useful. We have the same challenges of governance, federation and lifecycle management. No surprises here.

But there are also new challenges. The audience/buyers/drivers of Cloud are more midsize organizations that gain more benefit from Cloud computing. The time to integration is much faster and moves from weeks or months to minutes or days. There is an expectation that coding is no longer necessary and that self-service is available and straight-forward, performed by “citizen integrators” rather than a horde of highly technical development teams.

Closing out, Pezzini pointed out the following: There is an expectation for a single data/application/process integration platform. Having two platforms to move batch and real-time data is no longer acceptable. Governance is asynchronous and is decided by the vendor, not by your organization. Security is a major concern, but believe it or not, is likely stronger in Cloud than in the average enterprise.

iPaaS and Cloud integration

iPaaS may not mean the same thing to everyone, but to Pezzini, it means a hybrid combination of ESB, data integration, B2B gateway software, MFT and Governance/API management. It comes as a service and without needing IT operations, is model-driven, packaged, supports self-service, collaborative and has subscription-based pricing. This is Cloud sourcing of integration services. Many adjectives, but a very clear picture.

Many classic integration features like storage of artifacts, administration and monitoring of interfaces, SDK’s all run in a multi-tenant Cloud environment. The only thing you need to provide are the many-to-many interfaces you’ll run on top of iPaaS. IPaaS isn’t only about Cloud services integration, it is also a key way to perform B2B integrations, though maybe not as common.

Before you think that’s it, two emerging capabilities are Web API publishing to enable customers or business partners to develop mobile applications. There may be no other way to support scalability issues when so many end-points consume services. iPaaS on a Cloud infrastructure solves both security and scalability. Lastly, many mid-sized organizations are using iPaaS to skip over the integration of traditional ESB rollouts.

Flexible deployment topologies exist for iPaaS Cloud services integration. It can be iPaaS in the Cloud, distributed iPaas and private iPaaS. The flexibility will accelerate the adoption of iPaaS. By 2016, Gartner expects one third of large and mid-sized organizations to embrace iPaaS adoption. That’s an enormous market. The growth will be driven by:

• Mid-sized enterprises choosing iPaaS over traditional middleware
• Departments, LOB’s and subsidiaries of large corporations will find it faster and maybe cheaper to adopt iPaaS as an alternative to the SOA backplane
• Increasing B2B requirements and a need to reduce cost and time
• iPaaS as a component of SaaS. This embedding forces iPaaS into the marketplace
• Large vendors like TIBCO that are embracing iPaaS and will seed skills and best practices in the marketplace

These are all very logical expectations and strongly support Pezzini’s assertion that iPaaS is real.

Choosing iPaaS, middleware or outsourcing

So how does an organization decide when to use their choice of iPaaS, middleware and outsourcing. Pezzini believes the first question is whether an organization wants to do it all themselves or is willing to outsource integration. Doing it yourself involves point-to-point integration or an enterprise service bus. But moving toward outsourced capabilities moves naturally to iPaaS or an integration brokerage, where someone else is chosen to be a service provider that takes full responsibility for the integration of all entities.

• Why point-to-point? Easy, cheap, throw-away. It exists more today than some may realize. As the interfaces grow, however, the cost escalates for maintenance, upgrade and change. This can be a dangerous direction for most organizations.
• Why ESB? This is a familiar approach that leverages what you already own. Monitoring, management and governance is straight-forward and known. But this could be expensive overkill for simple requirements with expensive skills sets.
• Why iPaaS? Rapid integration flows development, templates and community. This comes with no IT ops burden and can be funded as an operational expense instead of capital expenditure (much easier approval an ROI calculation). But this could be difficult for demanding, highly scalable requirements. This is a solution where Cloud services prevail. No surprise.
• Why integration brokerage? These are often fully offshore teams running on someone else’s platform. Subscription is simplified and there’s ‘one throat to choke’. However, control is surrendered and the relationship with the integration brokerage vendor is critical. Like many outsourcing schemes, there is a loss of control over the integration intellectual property. An organization needs to accept this as a long-term arrangement. This ends up being a solution with strong arguments for B2B integration.

The reality is that enterprises will have combinations of these environments while they move toward iPaaS and Cloud integration services. Governance is difficult and the CIO is truly a Chief Integration Officer.

Pezzini recommends a short-term approach of Cloud requirements, a 90-day goal of a proof of concept for iPaaS, and a longer-term approach of creating a plan for a holistic integration strategy. You have your homework assignment.

Google+

Christopher Taylor