CCGrid 2005 Recap: The Closest Thing to Being There

By By Omer Rana, Cardiff University, et al

May 23, 2005

The following article was written by: Omer Rana, Linda Wilson, David Walker, John Oliver, Ali Shaikh Ali, Simone  Ludwig, Ian Wootten of Cardiff University (United Kingdom); Adarsh Patil of University College (Cork, Ireland);  and Brian Foley and Ligang He of Warwick University (United Kingdom).

The 5th IEEE Symposium on Cluster Computing and the Grid (CCGrid 2005) was held in Cardiff, Wales. The event attracted over 300 delegates from the United States, Europe and the Asia-Pacific region. It featured four tutorials, over 75 papers in the main track, nine workshops and a range of invited talks. The opening speech was given by the Rt Hon Rhodri Morgan (First Minister for Wales), who welcomed the delegates and emphasized the importance of knowledge intensive industries and services as exemplified through e-Science. The conference also featured 17 posters, on topics ranging from Grid middleware to experiments over Grid infrastructure. Keynote talks were delivered by Carl Kesselman (USC/ISI) on the recently released Globus Toolkit 4.0 (GT4), and by Tony Hey (EPSRC) on the UK e-Science Program. Carl Kesselman noted, in particular, changes in the security infrastructure within GT4. Other invited talks were by David Pearson (Oracle) on the Enterprise Grid Alliance, by Stuart Schechter (MIT Lincoln Lab) on SSH/Cluster security, and Heiko Ludwig (IBM) on WS-Agreement. The use of WS-Agreement to support contract management was particularly highlighted by Ludwig, and examples of the use of WS-Agreement were provided. Additional informational talks were provided by Jorge Gasos on the European Grid Program, and by Peter McBurney on the European AgentLink III Network. A “Work-in-Progress” session was initiated at this event by Mark Baker (Portsmouth University) and Daniel Katz (JPL/Caltech), and is likely to remain an important part of this conference in the future. A short course on the use of the Triana toolkit for distributed workflow was offered to the delegates.

Although an attempt was made to ensure that both Grid and cluster computing were equally represented at the event, Grid computing dominated most of the sessions. It was useful to see that a number of Grid projects which were initiated at the start of this conference series are now beginning to deliver results. There was also interest in trying to integrate ideas from cluster and Grid computing, especially in trying to extend operating system concepts to larger-scale distributed environments. Based on the presentations, this appears to be a promising area of work, and many researchers are now beginning to discuss the general notion of a “Grid Operating System.” This theme also received prominence in the talk of Jorge Gasos from the European Commission. One aspect presented at the conference was the capability to federate computational clusters running MOSIX. The work provided useful insights into techniques for developing a campus wide Grid — consisting of clusters from different departments running the MOSIX system.

The applications workshops attracted significant interest and involved presentations discussing updates to, and application specific requirements of, Grid infrastructure. The types of applications ranged from support for electronic learning (e-learning) to bio-informatics and other applications of health care. There was a significant overlap in the papers presented at the electronic learning workshop and themes addressed at the Semantic Grid workshop. The e-learning workshop also included papers discussing experience of teaching Grid computing courses to students — and best practice that could be more widely applied. Providing business models for Grid services remains an important concern, and mechanisms for charging for such services were considered in the Grid Economics workshop. This brought together participants from the computer science and economics community to address issues related to electronic contracts and managing strategies for allocating Grid resources in some fair manner. This issue was also covered through papers in the main track of the conference. An excellent tutorial in this area was also delivered by Rajkumar Buyya from Melbourne University. The tutorial provided an overview of Grid Computing and mechanisms supported for job management based on an economic model. There was also discussion of major industry players working in the area. The tutorial discussed the GridBus, Nimrod-G and the GridSim Grid Market Directory systems. A very comprehensive coverage was provided, with particular focus on how economic models could be used to support scheduling constrained by budget.

The workshop on Cluster Security (chaired by Bill Yurcik) featured papers addressing disk management, network management and “rejuvenation” mechanisms. The last of these is a particularly interesting theme which demonstrates how a cluster can be recovered after a failure has taken place. The authors also presented a model for predicting faults within a cluster that could be used as the basis for a rejuvenation strategy. Presentations and discussions about cluster infrastructure issues associated with opening up particular ports on a cluster to allow job submissions and the use of specialist scheduling systems also formed an important part of the workshop. The enthusiasm and involvement of the workshop chair led to a very successful event, and one likely to continue at future CCGrid conferences.

The Peer-2-Peer workshop received the largest number of submissions of all the workshops. A variety of topics were considered, ranging from efficient routing mechanisms, resource discovery techniques and the important area of trust management. Providing incentive structures to allow users to share content in the P2P environment was also considered, with particular focus on specialist Grid nodes that host such services.

Running Data Grid applications such as High Energy Nuclear Physics (HENP) and weather modelling experiments involves working with huge data sets possibly of hundreds of terabytes to petabytes in size that are often kept over wide area networks. This can suffer from overheads introduced by cross domain connectivity, a variety of different types of middleware and unreliable infrastructure. This was referred to as “performability”; that is, the joint consideration of performance and dependability. This particular theme formed the basis of the 1st International Workshop on Grid Performability. The properties of performability that were investigated in the workshops were as follows: responsiveness, availability, utilization, integrity, throughput, accessibility, latency, reliability and privacy. The idea was that by considering the performability of the Grid systems, and hence providing predictions, administrators can achieve greater impact on system effectiveness and user satisfaction. An example of the work that was presented included discussion of an ontology for describing performance data of Grid workflows. The ontology used to describe a set of performance metrics are utilized for monitoring and analyzing the performance of Grid workflows.

The Work in Progress session involved discussion about registries, networks and messaging as well as workflow. The talks that formed the basis of this session were expected to have a longer-term focus, describing primarily issues that are of interest to the research community, but still at an early stage of exploration. The interaction between the audience, presenters and the organizers was essential to provide success for a session of this kind.

The Semantic Grid workshop also attracted a significant audience, and attempted to bring together individuals with expertise in Semantic Web technologies and the application sciences. Prominent in this particular session were the issues of preservation and representation that are becoming evident with use of Grid computing in scientific applications. Calls were made on a higher level of semantics for representing both data and services, to ease the problems of data harmonization such as scalability and data format differences which will result without them. In addition to this, general ideas were expressed with such infrastructures in place, such as the automated composition of workflows.

Grid applications are often involved with large volumes of data produced by data-intensive simulations and experiments on scientific instruments. In order to guarantee seamless automation and interoperation of distributed data, the need for adequate descriptions such as semantic-based data descriptions, models, services and systems becomes crucial. For example, the LHC (Large Hadron Collider) is expected to be operational by 2006-2007, with the LHC Grid functioning, to begin the production of simulation data consisting of Petabytes of data. Sciences such as biomedical science and bioinformatics produce smaller data sets but numerous, diverse and widely distributed files stored on individual desktops and databases.

Papers were presented in the areas of resource descriptions, semi-automatic preservation of scientific data, file-based data for Grid services and data integration techniques. One particular paper investigated data integration services in bioinformatics, whereby a semantic data access and integration service based on the Grid paradigm was introduced. This service uses ontologies for correlating different data sets and is a fundamental component of the ProGenGrid system. The ProGenGrid system is a Grid-enabled platform, which aims at the design and implementation of a virtual laboratory where e-scientists could simulate complex “in silico” experiments, composing analysis and visualization tools available as Web services into a workflow. A number of other papers were also presented on the general Grid-based workflow theme, and this remains an important area of concern to both infrastructure developers and application scientists. Presentations included discussions of specialist workflow specification languages (such as AGWL), the development of tools that allow monitoring of tasks within a workflow session, and tools that allow aggregation of services from different administrative domains (JISGA and WebCom-G). Many such workflow techniques are also contained in specialist Problem Solving Environments aimed at scientists within a particular application area. Applications considered included Computational Electromagnetics, real-time systems, and genome comparison. To date, there still does not appear to be any consensus on a particular workflow technique — and often it is difficult to combine workflow tools developed by different groups.

The use of a shared memory space (such as a tuple space) also remains an important topic within the Grid and cluster community. The workshop on Distributed Shared Memory (DSM) featured a number of papers discussing issues of supporting synchronization within a shared memory, issues of consistency management, and performance issues associated with the use of particular networking infrastructure (such as InfiniBand). The use of DSM to support “Mobile Grids” was a particularly interesting contribution. The authors discussed how the use of a Shared Virtual Memory paradigm could be viewed as an unstructured DSM. The SVM concept was then used to connect PDAs and similar mobile devices within a Grid.

Network and I/O management also remains an important topic at CCGrid, with a workshop and three sessions devoted to it. Issues discussed in the workshop and the main track included deployment of resources within a “Lambda” network, and in particular how lightpaths allocated to particular application streams could be more efficiently managed. Advanced reservation techniques to support a particular Quality of Service were also discussed. Requirements for a reliable multicast technique over a production network was discussed, along with an investigation of how this could be achieved using the NORM protocol family.  The authors also provided a useful discussion of existing Grid applications (mainly multimedia based) that require the use of multicast. Papers discussing the provision of efficient I/O over a cluster provided formed the other theme in the communications track. As part of the tutorial on high performance I/O, the presenter discussed file I/O at various levels of abstraction, including POSIX I/O, MPI file communication primitives, parallel file systems, and libraries that allow structured access to complex file formats. Some of the implementation tradeoffs, performance characteristics, and optimization work on these libraries were described. Since high performance computing covers a wide range of different kinds of networking and storage hardware, tools that work well on one system cause severe bottlenecks in others. Insight was given into when and where it is appropriate to use one level of abstraction over another, and how to provide “hints” to the various libraries to allow them to extract the best out of a given system.

An area that was not represented at CCGrid 2005 included work on sensor networks, and integration of these with Grid middleware. Although there is significant interest in this area (at the GGF and various research groups), especially in health care provision (http://www.healthgrid.org/), there was not significant representation of this theme.

This summary of CCGrid cannot fully cover the variety of work presented at the conference. An attempt is made to present some key themes at the conference, and to highlight some new and emerging areas of interest to the community. Interested readers are referred to the proceedings of CCGrid 2005 for additional information. The next CCGrid conference is scheduled to take place in Singapore in May 2006. CCGrid 2005 proceedings are available on CD-ROM for purchase, please contact [email protected] for more information.

Subscribe to HPCwire's Weekly Update!

Be the most informed person in the room! Stay ahead of the tech trends with industry updates delivered to you every week!

Q&A with Nvidia’s Chief of DGX Systems on the DGX-GB200 Rack-scale System

March 27, 2024

Pictures of Nvidia's new flagship mega-server, the DGX GB200, on the GTC show floor got favorable reactions on social media for the sheer amount of computing power it brings to artificial intelligence.  Nvidia's DGX Read more…

Call for Participation in Workshop on Potential NSF CISE Quantum Initiative

March 26, 2024

Editor’s Note: Next month there will be a workshop to discuss what a quantum initiative led by NSF’s Computer, Information Science and Engineering (CISE) directorate could entail. The details are posted below in a Ca Read more…

Waseda U. Researchers Reports New Quantum Algorithm for Speeding Optimization

March 25, 2024

Optimization problems cover a wide range of applications and are often cited as good candidates for quantum computing. However, the execution time for constrained combinatorial optimization applications on quantum device Read more…

NVLink: Faster Interconnects and Switches to Help Relieve Data Bottlenecks

March 25, 2024

Nvidia’s new Blackwell architecture may have stolen the show this week at the GPU Technology Conference in San Jose, California. But an emerging bottleneck at the network layer threatens to make bigger and brawnier pro Read more…

Who is David Blackwell?

March 22, 2024

During GTC24, co-founder and president of NVIDIA Jensen Huang unveiled the Blackwell GPU. This GPU itself is heavily optimized for AI work, boasting 192GB of HBM3E memory as well as the the ability to train 1 trillion pa Read more…

Nvidia Appoints Andy Grant as EMEA Director of Supercomputing, Higher Education, and AI

March 22, 2024

Nvidia recently appointed Andy Grant as Director, Supercomputing, Higher Education, and AI for Europe, the Middle East, and Africa (EMEA). With over 25 years of high-performance computing (HPC) experience, Grant brings a Read more…

Q&A with Nvidia’s Chief of DGX Systems on the DGX-GB200 Rack-scale System

March 27, 2024

Pictures of Nvidia's new flagship mega-server, the DGX GB200, on the GTC show floor got favorable reactions on social media for the sheer amount of computing po Read more…

NVLink: Faster Interconnects and Switches to Help Relieve Data Bottlenecks

March 25, 2024

Nvidia’s new Blackwell architecture may have stolen the show this week at the GPU Technology Conference in San Jose, California. But an emerging bottleneck at Read more…

Who is David Blackwell?

March 22, 2024

During GTC24, co-founder and president of NVIDIA Jensen Huang unveiled the Blackwell GPU. This GPU itself is heavily optimized for AI work, boasting 192GB of HB Read more…

Nvidia Looks to Accelerate GenAI Adoption with NIM

March 19, 2024

Today at the GPU Technology Conference, Nvidia launched a new offering aimed at helping customers quickly deploy their generative AI applications in a secure, s Read more…

The Generative AI Future Is Now, Nvidia’s Huang Says

March 19, 2024

We are in the early days of a transformative shift in how business gets done thanks to the advent of generative AI, according to Nvidia CEO and cofounder Jensen Read more…

Nvidia’s New Blackwell GPU Can Train AI Models with Trillions of Parameters

March 18, 2024

Nvidia's latest and fastest GPU, codenamed Blackwell, is here and will underpin the company's AI plans this year. The chip offers performance improvements from Read more…

Nvidia Showcases Quantum Cloud, Expanding Quantum Portfolio at GTC24

March 18, 2024

Nvidia’s barrage of quantum news at GTC24 this week includes new products, signature collaborations, and a new Nvidia Quantum Cloud for quantum developers. Wh Read more…

Houston We Have a Solution: Addressing the HPC and Tech Talent Gap

March 15, 2024

Generations of Houstonian teachers, counselors, and parents have either worked in the aerospace industry or know people who do - the prospect of entering the fi Read more…

Alibaba Shuts Down its Quantum Computing Effort

November 30, 2023

In case you missed it, China’s e-commerce giant Alibaba has shut down its quantum computing research effort. It’s not entirely clear what drove the change. Read more…

Nvidia H100: Are 550,000 GPUs Enough for This Year?

August 17, 2023

The GPU Squeeze continues to place a premium on Nvidia H100 GPUs. In a recent Financial Times article, Nvidia reports that it expects to ship 550,000 of its lat Read more…

Shutterstock 1285747942

AMD’s Horsepower-packed MI300X GPU Beats Nvidia’s Upcoming H200

December 7, 2023

AMD and Nvidia are locked in an AI performance battle – much like the gaming GPU performance clash the companies have waged for decades. AMD has claimed it Read more…

DoD Takes a Long View of Quantum Computing

December 19, 2023

Given the large sums tied to expensive weapon systems – think $100-million-plus per F-35 fighter – it’s easy to forget the U.S. Department of Defense is a Read more…

Synopsys Eats Ansys: Does HPC Get Indigestion?

February 8, 2024

Recently, it was announced that Synopsys is buying HPC tool developer Ansys. Started in Pittsburgh, Pa., in 1970 as Swanson Analysis Systems, Inc. (SASI) by John Swanson (and eventually renamed), Ansys serves the CAE (Computer Aided Engineering)/multiphysics engineering simulation market. Read more…

Choosing the Right GPU for LLM Inference and Training

December 11, 2023

Accelerating the training and inference processes of deep learning models is crucial for unleashing their true potential and NVIDIA GPUs have emerged as a game- Read more…

Intel’s Server and PC Chip Development Will Blur After 2025

January 15, 2024

Intel's dealing with much more than chip rivals breathing down its neck; it is simultaneously integrating a bevy of new technologies such as chiplets, artificia Read more…

Baidu Exits Quantum, Closely Following Alibaba’s Earlier Move

January 5, 2024

Reuters reported this week that Baidu, China’s giant e-commerce and services provider, is exiting the quantum computing development arena. Reuters reported � Read more…

Leading Solution Providers

Contributors

Comparing NVIDIA A100 and NVIDIA L40S: Which GPU is Ideal for AI and Graphics-Intensive Workloads?

October 30, 2023

With long lead times for the NVIDIA H100 and A100 GPUs, many organizations are looking at the new NVIDIA L40S GPU, which it’s a new GPU optimized for AI and g Read more…

Shutterstock 1179408610

Google Addresses the Mysteries of Its Hypercomputer 

December 28, 2023

When Google launched its Hypercomputer earlier this month (December 2023), the first reaction was, "Say what?" It turns out that the Hypercomputer is Google's t Read more…

AMD MI3000A

How AMD May Get Across the CUDA Moat

October 5, 2023

When discussing GenAI, the term "GPU" almost always enters the conversation and the topic often moves toward performance and access. Interestingly, the word "GPU" is assumed to mean "Nvidia" products. (As an aside, the popular Nvidia hardware used in GenAI are not technically... Read more…

Shutterstock 1606064203

Meta’s Zuckerberg Puts Its AI Future in the Hands of 600,000 GPUs

January 25, 2024

In under two minutes, Meta's CEO, Mark Zuckerberg, laid out the company's AI plans, which included a plan to build an artificial intelligence system with the eq Read more…

Google Introduces ‘Hypercomputer’ to Its AI Infrastructure

December 11, 2023

Google ran out of monikers to describe its new AI system released on December 7. Supercomputer perhaps wasn't an apt description, so it settled on Hypercomputer Read more…

China Is All In on a RISC-V Future

January 8, 2024

The state of RISC-V in China was discussed in a recent report released by the Jamestown Foundation, a Washington, D.C.-based think tank. The report, entitled "E Read more…

Intel Won’t Have a Xeon Max Chip with New Emerald Rapids CPU

December 14, 2023

As expected, Intel officially announced its 5th generation Xeon server chips codenamed Emerald Rapids at an event in New York City, where the focus was really o Read more…

IBM Quantum Summit: Two New QPUs, Upgraded Qiskit, 10-year Roadmap and More

December 4, 2023

IBM kicks off its annual Quantum Summit today and will announce a broad range of advances including its much-anticipated 1121-qubit Condor QPU, a smaller 133-qu Read more…

  • arrow
  • Click Here for More Headlines
  • arrow
HPCwire