Through the Mobile Lens: Cloud and Analytics on System z
Bringing everything together on System z gives organizations and users the best mobile experience
12/17/2014 5:29:36 AM |
By Richard Gamblin
Many organizations are experiencing great "eureka" moments when it comes to delivering rich, engaging mobile applications: Infrastructure matters.
As mobile users, we are incredibly discerning. We expect a level of sophistication and engagement from a mobile app that far exceeds the same from website. We expect the underpinning service to be at our convenience; that our sensor-laden smartphones enrich all aspects of the user experience; and for companies not just to protect our sensitive data, but fully leverage it to offer an unrivaled service, beautifully tailored to our individual needs.
So, why does infrastructure matter?
Simply put, it is because the infrastructure is currently the rate-limiting step in this mobile utopia. And it needn't be. An intelligent infrastructure will proactively support and enhance the mobile user experience—and do so efficiently and economically. Let's look at why all of these requirements make the current generation of IBM System z mainframes the perfect choice to host and deliver mobile services.
Scaling up for Mobile with Cloud
One of the organizations I work with operates a highly successful mobile service, currently reaching tens of millions of consumers. Next year they plan to take the service global with an addressable market of billions of consumers. While this is an extreme example, it neatly illustrates the uncertainty of scaling to meet mobile demand. Everyone wants to release the next super-successful app, but there is always the danger that it becomes one of the 80 percent that are used once and then deleted
The challenge: How to create a mobile service that scales to meet unpredictable demand. In this case, the mobile service may be the core business application, but this is increasingly the Mobile Enterprise Application Platform, such as IBM MobileFirst
A common approach is to create a system that starts small and modularly scales out to meet new workload demands. There are two challenges here. First, to be sure the initial demand is satisfied, an excess of infrastructure is put in place, so there is a costly up-front investment. This isn't too much of a problem, because the mobile app will (hopefully) be wildly successful and at some point that infrastructure will be filled. Which raises the second challenge: how quickly the mobile workload fills the hardware. Suppose it takes six months to fill; at the current rate of hardware improvements, waiting six months for the latest hardware breakthroughs would have given much better value for the investment.
Let's compare this view to the scale-up (rather than scale-out) environment of System z. In this case, it's still possible to start small—with just one processor. In fact, for organizations already running a mainframe, it's more typical to start with a fraction of a processor. From there, additional capacity can be incrementally added to meet emerging mobile demands. This can be added without pause or delay to the running mobile service, and it can scale from fractions of a processor to hundreds of processors, meaning that the infrastructure is only acquired at the time of need.
So far, so good, but typically the first response is: What about a cloud solution?
From a mobile perspective, cloud is the perfect solution here. It removes the problem of unknown demand and neatly wraps it up in a pay-per-use technology model that flexes up and down as needed. However, that mobile workload uncertainty cost is incurred by the cloud provider and, for many organizations who host their own on-premise cloud, this issue remains.
Suppose, however, there was a way to combine the incremental scaling capability of the mainframe, with the premise and pay-per-use model of cloud. This prospect of combining the two has proven extremely attractive, so much so that IBM recently introduced the Enterprise Cloud System (ECS)
. ECS provides a single, integrated system comprising a System z server, storage and cloud management software.
Exploiting these capabilities in ECS means that patterns for IBM MobileFirst Platform can be created and scaled to meet demand in seconds and minutes, rather than weeks and months. This offers a seamless mobile experience, available to the consumer at the swipe of a finger and delivered economically
at the convenience of the business provider.
Mainframes help deliver the must-have mobile experience that is always on. How does it enhance the user experience?
Mobile Gets Personal With Analytics
Many organizations are adopting analytics technologies to better understand their customers to provide a more personalized service, tailored to each individual's needs. This is particularly powerful when applied to mobile users. On average more than 90 percent of us keep our phones within arm’s reach, 24 hours a day. Organizations that use the contextual information from mobile and combine it with the insight from analytics can offer their products and services at the exact point we need them most.
Consider the retailer who identifies a customer entering a shopping center; a location-based message pushed to the customer's mobile phone at just the right time will see them take action in 75 percent of cases, perhaps bypassing a competitor's store
. Alternatively, think of the loan application started en route to the shopping center that isn't appropriate for an individual; rather than a simple decline, the bank could assess the customer's financial history and offer them a credit card instead, which could be signed off and collected in the local branch along the way. Both examples, combining mobile and analytics, enhance the customer experience and drive brand loyalty.
The challenge here is that the immediacy of mobile dictates the need for real time analytics. Waiting for an overnight analytical assessment of the data to then issue an offer is simply not fast enough; the mobile user needs it in the time it takes to swipe or tap through the app.
To genuinely obtain and use this insight in real time, the analytics engine must run alongside the online transaction processing systems. These systems, such as CICS and IMS, which process billions of transactions every day, sit alongside 80 percent of the world's corporate data on System z.
This combination of online transactions and rich source of data makes the mainframe the perfect place to house an analytics engine and gain the necessary insight to offer mobile users the kinds of services described above. In fact, the analytics technologies available on the platform can deeply exploit this mix; the IBM DB2 Analytics Accelerator
can now give answers to complex queries in seconds rather than hours; and the Signature Solution—Next Best Action
employs analytical models to offer that perfectly tailored service to each customer.
Thinking Back to Reach Ahead
It’s fascinating to observe that the business opportunities that are emerging today are placing challenges on IT systems that have been addressed one way or another during the last 50 years. It is because of this pedigree that the current generation of System z servers is uniquely placed to address the IT challenges posed by mobile users. Whether this is through a scalable cloud-in-a-box solution or insight through a centralized analytics engine, it is clear that the evolutionary path of the mainframe allows it to continually adapt to meet these new challenges as they arise.
Richard Gamblin is a Software Architect and IBM Technical Staff Member who advises organizations on the application of Mobile, WebSphere and System z technologies to deliver business value.