Think of Univa as the de facto implementational (or even the commercial) arm of Globus -- where Globus is the driving force behind today's grid technologies and the open source Globus Toolkit is the basis for nearly all of today's major grid initiatives.  This tight coupling between Univa and Globus is not an idle boast.  Along with Steve Tuecke another co-founder of Univa was Dr. Ian Foster, the creator of the Globus Project as well as its leading and guiding light.  And here is how Ian sees the relationship between Globus and Univa:

  "When Globus becomes the enterprise standard, Univa wants to be the reason why" 

That succinctly sums up what Univa is all about.

Formed in early 2004, and headquartered in Elmhurst, IL (with Argonne, IL being the cradle for grid computing), Univa's goal is to harness the power and potential of Globus grid technology in the form of enterprise-class solutions, products and professional services -- for commercial and government entities.

Steve, gives us an exclusive, in-depth look at what Univa is currently doing and its thinking in this interview with Anura Guruge, IT In-Depth's editor at large.

Steve Tuecke: The Bio

Steve co-founded the Globus Alliance, originally known as the Globus Project, with Dr. Ian Foster and Dr. Carl Kesselman. He was responsible for managing the architecture, design, and development of the Globus software, as well as the Grid and Web Services standards that underlie it.

He began his career in 1990 as a software engineer for Ian Foster in the mathematics and computer science division at Argonne National Laboratory. In 1995, Steve helped create the Distributed Systems Laboratory at Argonne which, under his management and technology leadership, became the premier Grid research and development group in the world. In 2001, Steve focused on Globus architecture and design, creating Grid and Web Services standards, and establishing the project's corporate relationships.

In 2002, Steve received Technology Review magazine's TR100 award, which recognized him as one of the world's top 100 young innovators. The same year, he also was named to Crain's Chicago Business "Forty Under 40" and described as one of the Chicago area's "best and brightest." In 2003, he was named (with Ian Foster and carl Kesselman), by InfoWorld magazine as one of its Top 10 Technology Innovators of the year.

Steve graduated summa cum laude with a B.A in mathematics and computer science from St. Olaf College.

Q1: Steve, to grab our readers attention straight off the bat, lets start with some specifics – in particular the open source, Globus-centric grid computing enabling product that Univa plans to roll out in the second half of 2005.  What platforms will it support to begin with and will it just be another tool-kit so to speak or will it support some specific mission-critical applications?

A1: Univa products will support various Linux distributions, proprietary UNIX-based platforms such as AIX, HP-UX and Solaris, as well as MS Windows – we are contributing significantly to porting the core components of the Globus Toolkit to Window to further drive market adoption.

      A key differentiator of Univa Globus Enterprise (Univa’s commercial product based on the open source Globus Toolkit) will be enhanced installation, configuration, monitoring and administration (or ICMA) capabilities – the Globus Toolkit is already in use within large enterprises but some of these areas are often identified as gaps in the Globus Toolkit that have hindered adoption. We expect that the availability of commercially supported product from Univa will provide companies with the level of robustness and manageability expected of an enterprise class software product.

Q2: Steve, given that the most amount of unused enterprise computing power and resources are at the desktop level, i.e. on PCs, rather than at the data center per se, what plans and strategies does Univa have to incisively address this issue, which I for one have started calling the PCgate of the IT industry?

A2: Cycle scavenging, as this practice is commonly referred to, is one application of Grid technologies. In addressing your question, it’s important to make the distinction between “Grid applications” and “Grid infrastructure”. When most people think of Grid today, they think of running applications faster by parallelizing them and allowing them to tap into unused capacity on desktops or servers scattered through out the enterprise - there are a number of companies out there that provide tools to enable this.

      To elaborate -- Our focus is on providing Grid infrastructure, which you can think of as a layer that sits between multiple Grid applications and the underlying IT resources (computing storage, network) and provides a consistent, standards-based mechanism for applications to interact with the resources, irrespective of their type or location. This facilitates the decoupling of applications and resources, and allows the execution environment to managed dynamically through the discovery, reservation, provisioning, monitoring and decommissioning process, while maintaining quality of service.

      The cycle scavenging example you mentioned is just one of many workload types that can be supported by this Grid infrastructure, but it is not an area that Univa is targeting. Other workload types include coarse and fine grained task scheduling for high performance computing applications (e.g. risk analysis, portfolio valuation, structural design analysis), and data driven workflows, where large amounts of distributed data need to be managed through various stages of computation (e.g. business intelligence applications). Enabling a mix of these workload types to run on a common set of resources with appropriate quality of service levels is what Univa (and Globus) is all about.

Q3: Steve, staying with the problem of the gross underutilization of desktop computing resources, i.e. the PCgate problem, does Univa, in its evangelical (and consulting) endeavors, ever advocate a move to the terminal server model as a quick and easy way to partially solve this problem – independent of grid computing?

A3: What you’re describing is essentially desktop consolidation, and there are certain environments in which it makes sense. There are different approaches to achieving this, e.g. Citrix, Windows Remote Desktop, etc. but all of them require a set of computing resources dedicated to this function in order to ensure that quality of service is maintained.

      We propose using the same Grid infrastructure described above to support this application, which allows the same set of resources to be used for desktop consolidation and other applications, while achieving the service levels required by each individual application.

Q4: Steve, when it comes to enterprise customers, how does Univa see utility computing complementing or competing with grid computing and should enterprises now be thinking of utility computing as a mid-term solution for some of their data center server needs?

A4: Utility computing is as much a business model as it is a set of technologies – we consider utility computing as an extension of the Grid computing model outside the business boundaries of the enterprise.

      We advocate that companies should focus initially on deploying the appropriate Grid infrastructure within the organization. Once these standard management interfaces and mechanisms are applied broadly and consistently across the enterprise IT landscape, it is relatively straightforward to extend the boundaries of the Grid infrastructure so that it includes an external service provider. Without this infrastructure in place, companies signing up for one of the current utility computing offerings are creating yet another silo, from an IT management perspective.

Q5: Steve, can you please share with us your favorite grid computing success story and it is OK to be ‘generic’ if you can’t divulge the customer name or application as yet?

A5: We do run internal We worked with a customer last year that was developing a data driven workflow application such as the one I described earlier. Their original roadmap for delivering a solution went out many months, given that they were planning to build most of it themselves. Using Globus components we were able to demonstrate the required functionality in a matter of a few weeks which translated to significantly reduced time-to-market with their product. This initial success has since generated interest from other parts of the organization to adopt Grid.

      We’re seeing this process repeat itself in the market as early adopters of the Globus Toolkit are starting to move from pilots and limited (single application or department) deployments to broader use of Grid infrastructure

Q6: Steve, for our last question, I have to ask you this because it is so germane, though I fully appreciate that might have to tap dance around it in order not to upset any of your Globus partners, but do you really feel that the hardware vendor community is serious about grid computing or are they still just paying lip service because their first priority is to sell more server capacity?

A6: We’ve seen a number of hardware vendors embracing Grid technologies, driven by customer demands to more effectively manage and utilize their existing resources. The larger vendors have implemented a solutions-based selling approach where hardware is only one component of what they deliver to the customer. The focus of sales efforts is shifting towards ensuring that customers have an IT environment that is more flexible and responsive to changing business needs – just “throwing more hardware at the problem” is not the answer in most scenarios. As a result of this approach, many customers are realizing that they can actually expand or transform their business capabilities through the use of Grid, which ultimately results in additional hardware and software being deployed.

Thank you very much Steve.

Steve this was a great honor and a pleasure.  You have a done wonderful job of very clearly articulating the answers to many of the real-world issues people have about grid computing.  You are a true evangelist for this technology and I, as an avid believer in the need for grid, thank you for all your efforts.  I have interviewed Ian too but have not had the chance to publish his views.  I hope to do that soon.  I really appreciate you taking the time to share all this information with the "IT In-Depth" readership -- many of who are evaluating how they might be able to use grid at their workplace.  So this was extremely topical and useful.  Best wishes and thank you again.