by David F. Carr
Talk with business leaders about behind-the-scenes cloud technologies such as platform as a service (PaaS), and watch their eyes glaze over. Yet PaaS—done right—solves some of today’s most-urgent business problems, such as how to build mobile apps more quickly, tap into the Internet of Things, or add that one special feature to a cloud app that customers are demanding.
PaaS is different from the cloud’s two other service models: software as a service (SaaS) and infrastructure as a service (IaaS). The business value of SaaS is easy enough to understand—application software as a subscription service accessed over the internet. IaaS is simply internet-accessed hardware resources such as storage and compute. PaaS is more abstract and sounds more technical.
So the key to conveying the true value of PaaS is to talk about business benefits. PaaS can, for example, enable a company to launch a new customer service by adding a new software feature quickly. In general, PaaS helps companies do just about everything they need to do in technology more quickly.
“We’re living in a world where technology all of a sudden can do more than the business can absorb,” says Holger Mueller, an analyst at Constellation Research. That’s a far cry from the precloud era, when technology was usually a limiting factor. “Now two people with a credit card can do things that would have taken 100 developers 5 or 10 years ago—and lots of money.”
With a PaaS environment, one that includes strong application building tools and ready access to the infrastructure to support scalable and reliable applications, businesses can be much braver about using technology to distinguish themselves from competitors. Mueller thinks the companies tearing up old business models and creating new ones will produce more software in the next 10 years than they did in the last 20 years.
But PaaS must deliver a lot more than APIs and lightweight integrations, which some cloud vendors offer. PaaS must encompass database, industry-standard development, integration, analytics, and more capabilities. It’s the full stack of technologies needed to build modern enterprise applications. In other words, we don’t just want APIs; we want a platform that will let us exploit APIs, deploy our own novel code, connect to legacy systems, and field applications that solve the problems of customers, partners, suppliers, and employees.
Which sorts of problems? Consider the following examples.
1. Go mobile in a hurry. Most businesses have only begun to tap the potential of mobile computing. The smart ones are thinking hard about fielding the right apps to improve customer loyalty and boost employee productivity.
“PaaS specifically and cloud back ends in general are a natural for mobile computing,” says Al Hilwa, program director for software development research at International Data Corporation (IDC). “In fact, it is no accident that mobile and the cloud are rising phenomena at the same time. Mobile apps need back ends, and mobile developers love the abstraction provided by PaaS.”
Mobile devices sit outside of the firewall, Hilwa notes, so even in enterprise settings it makes sense for those companies to do at least some of the processing in the cloud rather than in their own data centers.
PaaS speeds up mobile app development by automating many steps in the process. For example, Oracle Mobile Cloud Service shows prebuilt APIs and recommends the most likely data connections, which a developer can accept or reject.
A good PaaS also provides tools that vastly simplify the process of creating mobile apps and rapidly deploying them. “A business user with no programming knowledge can create a decent mobile app, one that works across platforms, something that used to be tremendously challenging,” Mueller says. Even if that app needs to be refined by a professional programmer and designer, that ability to rapidly prototype means the business gets exactly what the business needs—with no surprises, he says.
2. Add a critical function to a SaaS application. Cloud applications have a reputation for being “configurable but not customizable,” says Sona Manzo, vice president of human capital management solutions at Hitachi Consulting. Limiting customization can be a good thing because it steers organizations toward standardized processes and best practices. “But there are unique circumstances where a customer really does have something compelling they would like to address that isn’t yet in the suite, or frankly may never be because it’s a niche need,” Manzo says.
Hitachi discovered such a unique need in its Global Development Center operations in India—to get employees company-provided transportation when they’re working odd shifts, when safe and reliable transportation is hard to come by. Employees working with clients in other time zones often have to work nonstandard hours, and project team deadlines may require extended hours that rule out normally utilized public transportation options.
Hitachi customized its Oracle Human Capital Management dashboard to let employees request transportation, either for a one-off special circumstance or on an ongoing basis for a longer project. That request is routed through management reviews and, once approved, to a distribution team that schedules the transportation. “We added the functionality we needed, with the same look and feel and user experience the rest of the suite has,” Manzo says.
Whether in human capital management, customer relationship management, or any other enterprise application, the combination of SaaS and PaaS gives businesses simplified access to standardized software, with the option to tweak or extend it as necessary.
3. Tap into the Internet of Things. One of the most exciting frontiers in business and IT involves the analysis of data generated by smart, connected consumer and industry devices—the so-called Internet of Things (IoT).
Manzo foresees manufacturers using smart badges to record when factory workers arrive and leave a facility, making time clocks unnecessary. In the world of logistics, companies are instrumenting shipping containers, pallets, and sometimes individual product packages to report back their location and other data, such as temperature for a product that needs to be kept cold. Automobiles, home appliances, and industrial equipment are being instrumented to report diagnostic and maintenance information to their owners, their manufacturers, or both.
If the possibilities are endless, so is the potential for companies to get overwhelmed by the flood of data coming in. “You can’t even buy servers fast enough to keep up,” Constellation’s Mueller says, making the cloud the only practical environment for most of these applications.
Product managers and engineers—not IT pros—tend to drive the IoT discussion, and those people don’t have the time or technical knowledge to set up a back-end system, says Michael Ottoman, chief operating officer of mFrontiers, which sells an IoT platform that runs on top of Oracle Database Cloud. About 90 percent of companies coming to mFrontiers want a cloud-based platform.
IoT applications create a huge, often unpredictable new data stream to manage, making scalability a major concern. “IoT processing is data-intensive, and the processing of this data to extract from it nuggets of information, and to support important decisions, is often best done in the cloud,” IDC’s Hilwa says. Like mobile apps, most IoT apps go beyond the bounds of the enterprise as well. “Data born outside of the firewall is largely being processed outside of the firewall,” Hilwa says.
4. Connect applications to create a new business process. Line-of-business leaders have no trouble buying cloud applications without IT department involvement. The hard part comes when they want to do things such as revamp a whole business process, which means having to integrate that application with other applications that might sit on premises or in the cloud.
“SaaS applications and cloud-based functionality give you best of breed, but how do you bridge it into your existing ecosystem?” asks Faisal Ghadially, a technology advisor at PwC. Building the connectors between two systems would typically take several weeks; PaaS systems such as Oracle Integration Cloud Service now offer prebuilt connections that can cut that down to a few minutes, Ghadially said during a recent webcast on oracle.com.
Semiconductor company Global Foundries is looking to connect cloud apps from a variety of providers, including Oracle and Salesforce.com, and to link cloud apps to on-premises apps such as Oracle E-Business Suite and Oracle’s PeopleSoft applications. Today, Global Foundries has integration specialists who test and tweak its application connectors; with Oracle Integration Cloud Service, it hopes to let the application experts take over that integration work.
PaaS specifically and cloud back ends in general are a natural for mobile computing. it is no accident that mobile and the cloud are rising phenomena at the same time. . . . mobile developers love the abstraction provided by PaaS.”–Al Hilwa, Program Director for Software Development Research, International Data Corporation
Those application developers “really have the business knowledge and domain knowledge to understand what mapping we need to do from a business perspective,” said King Ou, global director of systems and solutions at Global Foundries, in the same webcast.
Ghadially sees another benefit once companies can tap integrations written and recommended by other PaaS users. “You now open up an entire community of developers, so your in-house IT doesn’t have to build it all,” Ghadially says. “It now can be anyone in the community.”
5. Quickly launch an app, while meeting compliance. Speed is one of the biggest reasons companies choose a cloud platform to develop and launch a new app. When they decide not to use a cloud platform, it’s usually because there’s some regulation or company policy prohibiting it.
Some PaaS platforms, however, support developing and testing an application on the platform and then moving the application and database to a company’s own data center. That portability story is stronger if the cloud and the corporate data center environment are functionally equivalent.
Oracle’s approach has been to offer cloud versions of all its core products, making it easier to move data between its database and its database cloud service or between Oracle WebLogic Server and its Java middleware in the cloud. “Oracle goes to great lengths to ensure that, because it has a number of customers in the regulatory space that need to do that,” Constellation’s Mueller says. “That’s fairly unique for Oracle.” Companies can even take advantage of a cloud service called Oracle Database Cloud Exadata Service that delivers Oracle’s scalable, engineered systems for running Oracle Database. The point is that for every cloud service, there is an on-premises equivalent to make it feasible to move workloads back and forth.
Even in the Oracle world, though, maintaining portability means paying attention to the details. For example, new product features often appear first in the cloud versions of Oracle products, even though they may be on the roadmap to be added to the on-premises equivalents, Mueller says. “You have to be very careful if you want to deploy back on premises that you don’t pick up dependencies on things that work only in the cloud,” he warns.
6. Make data visual. The big data created by mobile, social, and IoT applications is only as useful as it is understandable. It doesn’t make sense to upload massive quantities of corporate data from an on-premises system into the cloud for analysis, but it does make sense to use the cloud to analyze data born in the cloud. Plus, new cloud-based visualization tools, available as part of PaaS, can let analysts bring in only the data they need to research a problem and then use data to inform colleagues of what they’ve learned.
One reason for doing data analytics in the cloud is that it may provide more-advanced capabilities. For example, one way to make data more understandable and to uncover patterns within it is through visualization. Oracle Business Intelligence Cloud Service, part of the Oracle Cloud Platform, includes a key feature, Visual Analyzer, that’s not yet available in its on-premises software equivalent, Oracle Business Intelligence Enterprise Edition.
Even nontechnical users can click on a few related data sets, and Visual Analyzer will automatically generate what it thinks is the best visualization for that selection. “It does a pretty good job,” says Dan Vlamis, who with his brother Tim cowrote Data Visualization for Oracle Business Intelligence 11g (McGraw-Hill Education, 2015), and runs the Vlamis Software Solutions consulting practice. “Does it always do what we would think is most optimal? No. But it’s an exploration tool,” Vlamis says. Business users can continue to refine the visualization, either on their own or with the help of a professional analyst.
Meanwhile, Oracle has gotten much better at uncluttering its default data displays. “It’s very sophisticated, very progressive, and very much in keeping with visualization best practices, which wasn’t necessarily true of some of the legacy products,” Tim Vlamis says.
Oracle says an Oracle Business Intelligence Enterprise Edition version of Visual Analyzer is forthcoming, but it’s easier for Oracle to roll out new features such as this one first in a cloud environment, where it has complete control, the Vlamis brothers say. In contrast, a conventional enterprise software release must be tested across multiple hardware and operating system configurations, with install scripts and database administrator documentation added.From Technology to Business Case
So demand for PaaS is rising—except that PaaS per se generally is not what customers are asking for, says Accenture Managing Director Terri Strauss, who oversees a joint venture with Oracle focused on the potential of the cloud. “It’s not usually articulated that way,” Strauss says. “It’s articulated through the business need.” Business leaders might start with a cloud app, but then they want more than standard functionality. “We’re finding clients asking first for SaaS, and then extensions of SaaS—which is how we get to PaaS,” she says.
Can you launch a mobile app in time for an upcoming industry expo? Can you show the sales team inventory held in a legacy enterprise resource planning system? Can you help managers explore what might be causing a sales slump? These are all business problems tied to technology. And increasingly, IT will be turning to PaaS to quickly deliver the databases, integrations, development frameworks, and analytics that help solve these problems.
Photography by Shutterstock