GCJ: How does Univa fit into the Globus ecosystem?

Miller: Univa's role in the Globus ecosystem is three-part:

1 - Response to the market - Univa will offer Globus-based products and services to enterprise Grid customers (both corporate and government). The Globus Toolkit has become the de facto standard for Grid implementations - and Univa is responding to the market demand for a new class of "enterprise ready" products and services to further support these Globus infrastructures.

2 - Advocate for the enterprise customer within the Globus community - Univa is a major contributor to the open source corpus of Globus software, and actively participates in the community that's steering the related standards. In this way, we're a conduit that's helping to steer the Globus code and standards in the directions that best support enterprise customer's Grid environments.

3 - Defining Globus' application to mainstream enterprise environments - Univa is educating the enterprise about common IT scenarios and environments for which Grid (and the Globus Toolkit) is a practical solution. Historically, the Globus Toolkit has been marketed effectively within research and science arenas, but has not had the proper documentation, use scenarios, and general product marketing support that an enterprise solution should have. So there remain some great opportunities to raise the profile of the Globus Toolkit in the enterprise, and Univa intends to play a central role in that education process.

GCJ: The Globus Consortium recently tapped Univa for some Grid projects - tell us more about the work you've been chosen to do.

Miller: The first project is the creation of a process by which the technical issues (aka bugs) that are the most immediate and serious impediments to "production Globus" are addressed. The bugs are selected and prioritized by the community of technology vendors (the members of the Consortium), and then addressed by the Univa engineering personnel with the most experience and background in that aspect of Globus. This constitutes a "rapid response" process that's most attentive to the commercial interests, with the resulting solutions going immediately back into the open source corpus.

The second project recognizes Univa's expertise and focus on the practical aspects of developing Grid services standards within the larger community of requirements for Web services. The Globus community, and particularly the efforts of Steve Tuecke and Karl Czajkowski, has been a major contributor to Web services standards such as WSRF. The Consortium is supporting our efforts to actively participate and influence the evolution of Web services standards in a number of standards bodies. The resulting specifications will address the requirements not only of the parties looking to use WS-* for application integration and composite applications (i.e., Service Oriented Architecture), but also the creation and use of Grid services - and service-oriented infrastructures - based on the same standards.

The third project speaks to making Globus the "technology of choice" throughout the world - the globalization of Globus, so to speak. This effort is designed to return to the open source corpus of software components that are easily localized, deployed without a dependency on or "bias" towards a particular natural language, and without a dependency towards any particular national or regional operating system environments. This project is meant to reduce the impediments to using Globus or deploying it quickly and reliably, particularly in the markets that rely on a different language and regional variations on the operating systems, such as the Asian and mid-East markets.

GCJ: A common perception in enterprise seems to be that Grid is still far from being a mainstream solution. Could you specify a couple of common enterprise IT scenarios for which Univa thinks the Globus Toolkit is an appropriate solution today?

Miller: Let's first do some definitions. "Grid" means too many things to too many vendors and customers. I'd have to say that for those vendors and end-users who are focused on enhancing a specific application, the notion of a "grid-enabled application" is becoming reasonably well-accepted within industry sectors such as financial services, automotive manufacture, electronics (particularly, chip) manufacture, life sciences and a few others. While Grid-enabled apps have become increasingly popular, they can also lead to their own application silos - collections of resources dedicated to a restricted number of applications or applications which must be supported by a single application platform.

Globus is about "Grid infrastructure" -- the federation of a potentially very heterogeneous collection of compute, storage and communication resources in a way that also supports a very heterogeneous collection of applications. In other words, a Grid infrastructure opens up for the enterprise the ability to add new applications and new application platforms to the collection pretty easily, and permits the enterprise IT organization to set policies by which the allocation and re-allocation of these resources can be directed ... in part by the real-time demands being placed on the applications themselves, and in part on the relative priority that has been granted to the applications.

With respect to common enterprise IT requirements for which Grid is a practical solution today - one of the major appeals of Grid in enterprise today relates to the management and transfer of data; an issue at the top of every enterprise's priority list. Increasingly, because of requirements for backup, recovery and business continuity, enterprises must manage enormous collections of data, "herding" it around the country or around the world. In addition, compliance with today's regulation requires a company to save EVERYTHING in a manner that's recoverable, but also in a manner that meets stringent requirements regarding security and protection from intentional or inadvertent disclosure. Globus-based infrastructure can make these efforts less costly in terms of capital expenditure, and even more important, less costly in terms of the PEOPLE costs associated with the administration and management of this "data herding." In fact, one of our first customers and partners is incorporating Globus-based infrastructure. We've just announced our collaboration with Raytheon, which is developing a platform for "data herding" to be used by their customers, who are primarily government agencies.

Optimization is another reason why Grid is a compelling, practical solution for enterprises today - they're interested in reducing un-used compute, storage and communication resources. In many situations, the demand for "compute power" can vary during the course of a day or a week by orders of magnitude. Anyone who's tried to do their "web shopping" at the most popular internet sites the week before Christmas knows the result. Either the company has "over-provisioned" their infrastructure to be ready for the busiest, highest levels of demand, or they're "under-provisioned" and haven't made enough resources available. Globus-based Grid infrastructure is one way to allow this responsiveness to demand, bringing more compute power to bear when there's need, and de-commissioning that extra resource for use by some other IT process when its NOT needed. At the SAP TechEd (developer's) Conference, they demonstrated an application which provides the "back office" for internet shopping carts that had been configured to use Globus-based infrastructure. As more web-customers appeared, the application discovered, reserved and then "commissioned" additional computers from a pool of computers. When performance of the site could be maintained with fewer resources, the application "de-commissioned" the unneeded computers and returned them to the pool so that they could be used by other applications.

GCJ: How important will services and support be to the uptake of Grid in the enterprise? Why would someone tap a startup for Grid services work when there are other major systems vendors that have been building Grids in the commercial sector for years already?

Miller: Professional services will be very important to the uptake of Grid infrastructure. And, it's pretty easy to understand why. Grid infrastructure isn't something that comes ready-made. It's not a "turnkey solution," since almost every enterprise has a very unique collection of resources, a unique combination of applications that need to run on the infrastructure, and very different levels of performance or service levels that are acceptable. Thus, for the foreseeable future, most implementations of Grid infrastructure will be consultative sales, and require the expertise of professional services to design and then implement the Grid infrastructure. Having once put Grid infrastructure in place, any enterprise which depends on that infrastructure will also demand that there be a high-quality level of technical support. No one wants to "bet the company" on an infrastructure for which you can't find a "mechanic."

As for your second question: Why would someone tap Univa, a startup, for Globus-oriented services and support? Well, the founders of Univa are truly the inventors and the world experts in Globus. Working for our company are perhaps the most skilled people on the planet when it comes to understanding what Globus can and cannot do. We envision Univa's professional services to be sought out by leading enterprise users of Globus, particularly for the assessment, design, customization and integration of Globus components. We also expect to partner with the major systems vendors and professional services companies in executing and supporting the infrastructure for major projects.

GCJ: Univa has hinted that it intends to sell its own, enterprise-class variation of the Globus Toolkit software. What new features / capabilities is Univa integrating, and why will these enhancements be attractive to enterprise?

Miller: We aren't pre-announcing the features and functionality of Univa Globus Enterprise, but there are two aspects we should speak about. First, a point of clarification: Univa isn't developing its own variant of the Globus Toolkit. In other words, we will not be creating a competing version of the open source GT 4. Rather, we will be delivering to our customers vetted, tested binaries (compiled versions) of selected Globus components, which are tested out and designated as "ready" for use on specific operating systems or to be integrated with specific compute, storage, network or application resources. Anything that has been written to use the open, published Globus interfaces should run as expected on UGE or on a correctly compiled, installed and configured open source Globus component. And, that brings me to the second point.

What UGE also includes is a set of added-value offerings that assist the developer and IT manager to install, configure, manage and administer Univa's Globus components. Finally, we expect to provide the "connectors" or integrations between Globus and a wider range of commercial resources than will be found in the open source versions.

I believe that the attractiveness of this approach will be pretty obvious. We're compatible with any use of the open standard or openly published Globus interfaces, but with UGE, we have reduced the effort involved in deploying Globus and integrating it into the enterprise's IT environments.

GCJ: Does Univa plan to do direct sales only, or is it also pursuing a channel strategy?

Miller: Great question. In fact, we believe that Grid infrastructure is mostly to be sold in conjunction with other infrastructure resources, or by large professional services organizations with broad practices in Grid and the related infrastructures. Thus, we are placing a lot of emphasis on a channel strategy.

We do believe that Univa's professional services group, which will be engaged in some direct sales, will be a good source of technology sales, but that this level of sale will be significantly overshadowed by the sales of UGE through our channel partners.

close window