RealTime IT News

Mark Linesch, Chairman, Global Grid Forum

Mark LineschGrid computing is in a tight spot these days.

Vendors such as IBM, Sun Microsystems, Oracle and others are throwing their weight behind initiatives that allow many computers work to solve complex problems, which is a good thing.

But the licensing issues that face grid computing vendors threaten to thwart their ability to release products to customers without ripping them off or giving away IT resources for a song.

Enter Mark Linesch, chairman of the Global Grid Forum (GGF), a not-for-profit group of users, developers and vendors working toward the global standardization and adoption of grid computing.

Before joining GGF, Linesch cut his teeth as vice president for Hewlett-Packard's Adaptive Enterprise Program working on grid and distributed computing architectures.

It is exactly because GGF is a standards body that the public doesn't hear a whole lot about what it does. Looking to fix that issue, Linesch discussed the state of grid computing ahead of the GridWorld show, slated for Oct. 3 to Oct. 6 in Boston.

Q: Companies in financial, manufacturing and life sciences sectors appear to be ahead of others in adopting grid computing, but it seems that commercial grid applications have been evolving. Talk about the evolution from grid computing in the research arena to the enterprise.

Grids (and GGF) grew out of the high-performance computing industry where big compute problems like risk analysis in the finance sector and drug discovery in the life science sector were prevalent.

We see the same type of hunger for compute cycles, massive amounts of storage, and network bandwidth in engineering design and in many hard sciences, such as astronomy and high-energy physics.

In the past two years, enterprise case studies have revealed the impact grids are making on other departments, such as customer service and portfolio analysis.

For example, quotes or analyses that previously took 15 to 30 minutes to complete can now be done while the customer is on the phone. There are many more examples of how grids are enabling enterprises to adjust their business processes, reduce cycle time, better meet the needs of their customers and gain competitive advantage.

Q: How are telcos going to contribute to the current grid and utility computing boom, if at all?

We're still at the early stages, but GGF sees telcos playing a unique role in the future of grids for a number of reasons.

First of all, telcos provide connectivity to a broad range of enterprises and their IT resources, including computers, storage and even applications. Not all of these resources are fully employed, and enterprises may wish to receive a return on their investment by making these resources available.

Particularly, GGF sees telcos serving as resource brokers, fulfilling end-user requests for available resources through existing network bandwidth and switching ability.

Q: What, if any, are the major concerns about or shortfalls of securing grid architectures? How can vendors assuage customer fears about grid security?

What's unique to grids is that they center around the scalable sharing and management of distributed resources -- computing or storage resources, large databases, unique instruments and so on. In grid security, it is not the occasional human intrusion that people worry about; it is losing control over all those resources at the same time.

One of the key issues that GGF is currently working on is emergency response and distributed, cross-organizational authorization. How to quickly disable in a controlled fashion what the grid once enabled. We also look at how to best leverage the new standards and technologies that are emerging, such as the Web services specifications described in our WS-Security Profile now being adopted by the industry.

Q: We hear a lot of gripes from analyst firms about the problems with coming up with equitable software licensing schemes for grids. Is this something the GGF is looking to address? If so, what solutions do you have in mind?

Software licensing is just one of a number of issues which early adopters of grids are dealing with. In the case of licensing, while there are a few technical issues, most of the barriers involve non-technical, economic and policy issues that the industry must sort through. This is not meant to trivialize the issue, as it is clear that a strict "one license per CPU" model does not work well in a virtual collection of potentially hundreds of CPUs.

GGF provides a forum in which end-users, software vendors, and others can work collaboratively to identify reasonable and equitable options for all parties.

GGF is opening the door to explore new licensing models such as those that would charge for usage of the application, not just for the absolute number of CPUs available in the grid.

These time-based usage models incorporate rolling over minutes not used to a later time or not counting minutes used to review execution results. Clearly, there will need to be some flexibility and a lot of dialog over this issue.