The consumerization of the enterprise is driving tremendous change in all organizations. Our employees are bringing their own technology to work. Even more importantly, our customers are expecting us to be available to them on their smartphones and other devices, in their social networks, in online chat rooms and instant messaging sessions, and more.
The user experience is indeed changing, both for customers and employees who work at the point of customer contact. Changing the user experience in your organization — and across your customer- and employee-facing systems — has to be done right, or your efforts won’t live up to expectations. For example, in research conducted for Unisys, IDC projects that workers’ increasing use of consumer devices and social technologies will dramatically increase the number of business interactions. Those will drive a higher load on enterprise servers. Modernizing the front end without scaling the back end is a recipe for disappointment at best and failure at worst.
This is a driver behind our announcement today of some important — dare I say “breakthrough”? — new virtualization technology, dubbed sPar, for secure partitioning. We’ve launched the sPar technology along with some new models in our ClearPath mainframe line. Taken in total, our newest offerings can efficiently boost the performance of your IT resources, making it easier to modernize strategic application environments and respond swiftly to business changes, such as the consumerization of IT.
About three years ago, we recognized that the advent of consumer devices, Web applications and services, and cloud computing were going to fundamentally change how people interact with mission-critical IT solutions. This is not just an issue of the young crowd versus the old crowd. We believed this was the beginning of a major shift in how people work and drive organizational productivity.
Fast forward to 2010. Consumerization has taken hold. Everybody is doing it. The change is massive, pervasive, and still unfolding. It presents a rare opportunity for companies to serve customers and workers better than ever before. Yet it also presents exceptional risk — not just in security, but also in system reliability, scalability, availability, and latency.
Consider this example. In the hours before a flight, passengers today turn to their smartphones to check the airplane seating chart, and see whether they can improve their seat assignment. The passengers love this, because they can manage their travel experience with a handy device and interface they’re familiar with. The airlines love it, because self-service mobile apps reduce or eliminate the cost associated with having customer service agents perform seating changes.
Consumer technology, such as the iPad and the other new tablets that will soon flood the market, can bring similar benefits when put in the hands of customer-facing employees. Using the airline example again, let’s say a flight is canceled. Instead of dealing with a massive queue in front of one or two ticket agents, the airline can bring in supervisors and other employees.
Armed with their own wireless tablets, the airline employees begin rerouting passengers to other flights. Instead of a tedious first-in, first-out queue, the airline dispatches passengers asymmetrically. No additional kiosks are required. Everything can take place within the confines of the gate. We speed the process, and we reduce passengers’ frustration. And we save money because we’re getting people booked and on their way faster.
This type of thinking can be applied to almost any industry: public sector, insurance, automotive, legal, law enforcement, finance, military, you name it. Workers could travel the field equipped with commodity wireless tablets; meet with customers and colleagues; and enter, update, and retrieve information, including audio, photos, and videos captured with a tablet device. The tablet they would carry with them would provide standardized forms and other media served directly from their organization’s data center.
The lightweight tablet could be charged in their vehicle as they travel from point to point and would run for house on a charge. It could use the mobile phone network, so the workers would have wireless access everywhere. It would be integrated with the public Web, so they could tap other information, resources, and communication tools (including two-way video chat) as needed. It would let them be more productive with greater accuracy. It would provide a better customer experience and make for happier employees.
But behind it all you need a strong data center …
All of these scenarios are exciting, and appear simple at first blush. Yet delivering this experience requires a sophisticated pairing of consumer devices with robust, sophisticated data center resources. Smart wireless devices, custom mobile pages or even apps for each device OS and secure wireless networks require services-driven interfaces between the new devices and the data center’s mainframes, and a mission-critical data center architecture to support the load. If you don’t align the consumerized experience with a back-end system capable of sustaining the transaction and interaction loads that typically accompany these new generation applications, all bets are off.
Here’s a real-world example. One of our U.S. customers recently deployed a new ClearPath system for their sales organization, comprising independent agents. As these agents discovered the faster response time of the new system, their behavior and usage patterns changed. On the old system, they typically ran one carefully formulated query at a time and waited for a result. On the new system, response time was so good that they initiated many more requests to get a better picture of the options available to their clients. The result was that the number of server interactions per agent rocketed upwards.
The improved user experience opened the door to new user behavior. When you provide more responsiveness, a better environment, and a better customer experience, users just want more, and take more. Is that bad? No. But it means you have to plan your enterprise server strategy to anticipate and handle the extra interaction load. Likewise, here at Unisys, these are some of the issues that we have been thinking about and working on from an R&D perspective. And this is what makes today’s announcement of sPar so important. We have been working to help our clients prepare their ClearPath environments for the increased demands that go hand-in-hand with the consumerization of IT.
One successful approach has been the use of specialty engines — special-purpose processors that perform specific application functions — to supplement the central processing complex with specialized processing resources. The ClearPath ePortal, for example, lets workers using devices such as iPhones and iPads tap their ClearPath solutions through the industry-standard Web Kit, which delivers a familiar look and feel. Until today, these specialty engines were physical servers that were integrated with the ClearPath environment through high-speed connections.
Today’s announcement of the ClearPath Libra 4100 Series introduces an additional option for specialty engines through secure partitioning (sPar). sPar is a Unisys-developed, enterprise-class virtualization technology that puts specialty engines on the same physical platform as the operating system — initially ClearPath MCP. Equally important, they are managed under the direct control of the optimized MCP. This means they interact faster, are more easily managed, and are more cost-effective for the organization.
These virtualized specialty engines perform exactly the same functions as before. But instead of being separate interconnected servers, they’re seamlessly integrated with the same server that MCP is on. And the MCP, instead of communicating through high-speed external connections, now communicates directly through memory queues to the specialty engines. Execution at memory speed, as you know, is faster than the fastest LAN connection.
And that data center demands enterprise-class virtualization …
The sPar virtualization technology is a major advance for ClearPath. It is unique to Unisys, and is fundamentally different from the commodity virtualization technology deployed in the enterprise today.
In commodity virtualization, you take a platform and carve it up into smaller slices, called virtual partitions. Each partition behaves like a physical server, but is really an independent slice contained in a larger system. You typically have multiple partitions sharing the same processors and other resources. The partitions typically depend on a hypervisor to figure out how to time-slice the processor and other resources among them.
That means the partitions are also sharing I/O channels, arbitrating among themselves to get to the I/O services required. The end result is unpredictable and sometimes unreliable performance. When you run an application in a partition using commodity virtualization, you can’t know for certain if you’re going to get the right amount of processor or I/O resources at any given point in time to support a mission-critical application in a consistent fashion.
With Unisys sPar technology, partitions have whole processors in a multi-socket ClearPath system dedicated exclusively to them. In the case of the MCP partition, there would be multiple processors dedicated exclusively for MCP. There would be a different set of processors dedicated to the ClearPath ePortal function. No other partitions or applications use them.
The same holds true for the I/O channels, whether for interfacing to storage or interacting with the network connections or other servers. Those connections are wholly dedicated to the MCP or to ClearPath ePortal. There’s never any conflict. The end result: system performance for mission-critical requirements is consistent, which is the prerequisite for any enterprise-class solution.
Predictability is the first difference between commodity virtualization and sPar’s enterprise-class virtualization. The second difference is in management. Again, commodity virtualization carves a physical server into multiple independent virtual partitions. This translates into tremendous cost savings when compared to having to maintain one server per environment. That’s the upside. The downside: each of those partitions is independently managed.
If you take a server and put 30 virtual Windows Server 2008 environments on it to accommodate 30 different departmental applications, you are going to have to manage each one independently. Things like updating security patches, physical operations for the server, upgrading to a new release, tweaks, and so on all have to be done individually to each virtual server instance. There are some independent management products that can simplify the process, but something still needs to manage all of these operating environments.
With Unisys specialty engines, supported by sPar or external integration, management of the partitions is handled through the ClearPath operating environment. You achieve the benefits of enterprise virtualization through sPar and centralized management of the environment through the specialty engine implementation. You don’t manage the Windows environment. You manage the function.
As an example, ClearPath ePortal delivers Web services and smartphone integration, and it is all managed by MCP. Functions like security patches are managed through MCP. So management is much easier and more cost-effective, because it’s centralized into an enterprise-class environment.
The third difference between commodity virtualization and enterprise virtualization? Resiliency and reliability. In a commodity environment, in many cases, a failure in one virtual partition can cause other virtual partitions to fail. That does not occur with sPar environments. sPar has the equivalent of physically separate processing environments — that is, the same as if they were running in physically separate servers. If one virtual partition goes down for some reason, it doesn’t affect any others. sPar can maintain a much higher level of availability than you would find in a commodity environment.
That said, I don’t want to my comments to be misinterpreted. There are certainly scenarios where commodity virtualization makes sense and should be used. The areas are those associated with general-purpose requirements, such as departmental applications, and other non-critical functions can often benefit from commodity virtualization as well. But if you have a mission-critical or enterprise-class function that requires the highest levels of scalability, consistency of responsiveness, and manageability, then commodity virtualization simply won’t do.
We’ll be sharing more about sPar virtualization for ClearPath systems based on Intel processors in several webinars. I encourage you to sign up and learn more about this innovation, particularly those that deal with the consumerization of IT and its impact on enterprise computing systems.
Web Services for ClearPath MCP and OS 2200 applications – ClearPath ePortal
Wednesday, Oct. 27, 2010 11 a.m. – 12 p.m. EDT
Smart Phone integration with ClearPath OS 2200 and MCP applications
Wednesday, Nov. 3, 2010 11 a.m – 12 p.m. EDT