April 14, 2011
In scanning the floor during the GlobusWORLD conference event this week, it wasn't hard to miss Steve Tuecke. Not just because of he towered over most of the attendees or the fact that he seemed to be everywhere at once--there was something else--you could say that he seemed to be bursting with energy. This was probably due to his excitement to showcase a new extension of his long-standing work with the Globus Project in the form of a web-based service aimed at researchers with big data demands called Globus Online.
Tuecke has been a fixture at the Computation Institute in Chicago, Illinois for around 20 years and currently serves as Deputy Director. Many of the projects he manages touch the high-performance and distributed computing spheres in addition to his work in biomedical informatics.
As one of the co-founders of the Globus Project, along with grid luminary, Ian Foster, Tuecke has overseen core elements of architecture, design and development. More recently, he has been working with grid and cloud services that underpin Globus Online, which provides a robust SaaS tool that is set to solve one of the most troublesome issues for researchers—long or complex file transfers to and from distributed resources.
Although Tuecke explains some of the core concepts behind Globus Online (as do a number of others who will appear in our video interview series this week) in essence, this addition to Globus development efforts makes use of the software-as-a-service model via Amazon Web Services infrastructure to deliver a highly capable high-performance file transfer service.
While the Globus Toolkit was at the heart of any number of discussions during the GlobusWORLD event, the star of the show this year was definitely the Globus Online service. Proving that folks will do anything to win free Apple products and that the service does what it claims, a challenge was issued at the beginning of the conference for users to test the system. Markus Binsteiner, the force behind an interesting use case for Globus called Grisu, moved 228.13 TB and 533,062 files in 17 hours. (And no, he didn't pack it onto drives and ship it via overnight express.)
Users like Binsteiner and others remarked on the fact that Globus Online has every hallmark that is increasingly making software-as-a-service models attractive for scientific users (ease of use, reduction of complexity so research remains unobstructed by IT-related hassles, avoidance of research delays via much slower file transfer mechanisms, etc.).
According to a number of users I spoke with during the conference, including those from NERSC and elsewhere (as well as in the genomic sequencing context as discussed in our video interview with researcher Brigitte Raumann) this is quite revolutionary. Many of the scientists were using much slower data transfer mechanisms either via the Globus Toolkit or even via postal services that transferred vast data sets via packages. In both cases, there were some serious risks of data being lost in either mode of transport.
Steve Tuecke discussed a few of the challenges and benefits of Globus Online above, but there are some other topics he touched on that we should go back to for a moment.
While Tuecke provides a sound description of some of the evolutionary aspects of the grid to cloud transition (beyond the moniker matter) some of his more interesting points lie how he understands the benefits and drawbacks to cloud and grid computing models. If you happen to watch this all the way through you’ll get to the question about where cloud computing has succeeded commercially whereas grid didn’t have quite the same level of mass success.
His answer is worth repeating; although he says he hopes that much of the work on the grid side has allowed cloud computing and distributed computing in general to thrive, there was a convergence of some important elements that permitted clouds to take off. Among such elements are, of course, the fundamental technologies that back both grids and clouds. Certainly, back in 1995 when the Internet was still in its early stages of evolution for commercial software purposes in particular, it would have been difficult to garner critical mass around something that was being used in scientific and academic circles and internally in some large enterprise organizations. Furthermore, thanks to the business models that Amazon helped to develop with pay-as-you-go access to what looks like unlimited resources, cloud computing was able to find a wide audience and user base under the umbrella (at least at first) of one key service provider.
One of the more compelling aspects of this convergence Tuecke describes is the “sociological” piece that really brought the software-as-a-service approach to the mainstream. He uses the example of Salesforce.com and its delivery of business-critical applications via a web service as the poster child but really, with the development of web-based applications from email to collaboration and social networks, this evolution has been a quiet, swift, but massive. People now are simply expecting that they should be able to have access to core business or research tools both on their local machines and in the cloud. This is evidenced by any number of “traditional” (meaning here installed locally) software companies scurrying about to make sure that they can fine-tune their delivery and licensing models to accommodate this newer but urgent need for application delivery options.
In general, by moving to the SaaS approach and new services and application delivery options the work behind Globus Online holds something else for Tuecke and the team—unlike the days with Globus Toolkit, this new service provides a chance for them to interact more with end users and base new developments on feedback and use cases. Whereas before with the Globus Toolkit, the team worked with what Tuecke describes as the “unsexy” infrastructure software that really couldn’t be demoed, the team can get a more hands-on feel for the capabilities and barriers with their development efforts.
And on a side note for any of you that are interested, Tuecke told me that he’s hiring. Seems like a pretty upbeat, pleasant and driven guy to work for and from my interactions during the conference, this describes the whole Globus team.
10/30/2013 | Cray, DDN, Mellanox, NetApp, ScaleMP, Supermicro, Xyratex | Creating data is easy… the challenge is getting it to the right place to make use of it. This paper discusses fresh solutions that can directly increase I/O efficiency, and the applications of these solutions to current, and new technology infrastructures.
10/01/2013 | IBM | A new trend is developing in the HPC space that is also affecting enterprise computing productivity with the arrival of “ultra-dense” hyper-scale servers.
Ken Claffey, SVP and General Manager at Xyratex, presents ClusterStor at the Vendor Showdown at ISC13 in Leipzig, Germany.
Join HPCwire Editor Nicole Hemsoth and Dr. David Bader from Georgia Tech as they take center stage on opening night at Atlanta's first Big Data Kick Off Week, filmed in front of a live audience. Nicole and David look at the evolution of HPC, today's big data challenges, discuss real world solutions, and reveal their predictions. Exactly what does the future holds for HPC?