Grid Technology and The Globus Project - Assignment Example

The Globus Project – Critical Analysis By March 16, 2006 Defining Grid Technology Lumb (2003) set out to provide a working definition as per the OGSAinfrastructure and has been provided such a definition by GGF’s OGSA Working Group as involving two major pieces to OGSA - core Grid components and a common Grid infrastructure. Core Grid components target functionalities like resource allocation and policy management. Built on top of the infrastructure piece, these components can be combined to build Grid applications and services. The common underpinning is the Open Grid Services Infrastructure or OGSI. OGSI results from Grid-motivated extensions to Web Services and specifies a Grid Service. This Grid Service removes the need for core Grid components to directly reference specific protocols like FTP, HTTP or LDAP. In the pre-OGSA world, this was common practice out of necessity. It handcuffed components to protocols making integration and extension labor intensive By understanding this very explanative definition, we take a look at one such open source technology website that has been working on further development of the Grid technology since 1997 called The Globus Project. A Critical Analysis of the Globus Project The Globus Toolkit (now in its 4th version) “is a fundamental enabling technology for the ‘Grid’, letting people share computing power, databases and other tools securely across corporate, institutional, and geographical boundaries without sacrificing local autonomy”. (Globus.org). The Globus Project Website Initially entry into the Globus Project website is virtually similar to many other open-source code websites in that the main purpose of the actual site is not to be graphically pleasing, but, rather the principle purpose of software development. This analysis of the site will show merit as to the benefits of using just the Globus Toolkit as a major development into the grid anatomy or could it perhaps be better served by alternate technologies such as .NET or CORBA. The Globus site and grid technologies are utilized by many high end performance sites mainly those that are in need of resources that cannot be handled by just one server, or one computer. Much like SETI’s @home project of using a grid made up of thousands, if not millions, of computers to evaluate millions upon millions of data modules, the grid concept is used in the same capacity. The toolkit is the most important functionality of using grid technology and this is evaluated here with the Globus Project website. As Grid technologies are “poised to move into commercial computing, standards become critical”. (Lumb 2003) and to compete against such powerhouse software companies such as Windows and Unix, it may prove to be a valuable need that is fulfilled where others fail, or come close to failing. The Globus Toolkit The bridging of the gap between the user desktop and the grids is a continual need to ensure that the open source community is preserved, this was the driving force for inception of the Toolkit. In 1995, a small team led by Ian Foster at Argonne created new protocols that allowed users to run applications on computers across the country. This successful experiment led to funding from the Defense Advanced Research Projects Agency (DARPA, and 1997 saw the first version of the Globus Toolkit, which was soon deployed across 80 sites worldwide. The U.S. Department of Energy (DOE) pioneered the application of grids to science research, the National Science Foundation (NSF) funded creation of the National Technology Grid to connect university scientists with high-end computers, and NASA started similar work on its Information Power Grid. (Globus.org) The following diagrams show how this transformation has moved during its initial inception in 1997, by such high level organizations such as NASA and the DOE. The Globus Toolkit version 4 includes “software for security, information infrastructure, resource management, data management, communication, fault detection and portability”. (Globus.org) Foster, Kessleman and Tuecke (2001) describe what the Toolkit is projected as its function and “has been designed to use (primarily) existing fabric components, including vendor-supplied protocols and interfaces.” It seems apparent just from the reading initially about the Toolkit, that if a Vendor does not satisfy the requirements of a fabric-level interface, functionality will be missing. Therefore, this is extremely proprietary in the absence of any vendor specific designed software. The Grid architecture is shown in the following figure outlining its relationship to the Internet protocol architecture. The initial development of the Grid system apparently was to satisfy the requirements of a community of developers that needed to access different resources on different types of community machines. From this next diagram you can appreciate the needs to utilize the resources others had on their computers without spending unessential software specific to their field of industry. .NET and CORBA Platforms This type of sharing relationship can also be found in .NET platforms with the new peer-to-peer networking functionality that can be built to satisfy this direct need to share resources that are found in the respective peer network in order to make available resources that are otherwise inaccessible to other group members on their own systems. Although the Globus Toolkit itself has proprietary software developed specifically for this toolkit by the developers of the toolkit software, it provides great limitations based on a “Generic Authorization and Access (GAA) control interface.” (Foster et al 2001) and doesn’t take into consideration the immense amount of security that is offered through the .NET platform for server deployment. It is an excellent concept (Globus Toolkit) but would seem to lose a lot in public acceptance and availability. Many of the collaborative sites that are used on the Internet would be well utilized by more robust software packages such as .NET or CORBA. Other than for scientific or logic reasoning collaboration, this may in fact have its limitations for general application. The author’s discussion surrounding the desirability of other communities or delivering functionality to other applications does not seem efficient in the sense that there may be a security infringement on other proprietary software such as Windows with respect to adding in the Toolkit’s own API’s and since the advent of new peer resource offerings from Microsoft, this software may in fact be relegated to a few interested community hosts. The authors of the article The Physiology of the Grid: An Open Grid Services Architecture for Distributed Systems Integration explain that the development of grid technologies was in answer to the need for “resource sharing within far-flung scientific collaborations. Applications include collaborative visualisation of large scientific datasets (pooling of expertise), distributed computing for computationally demanding data analysis and coupling of scientific instruments with remote computers and archives.” (Foster, Kesselman, Nick and Tuecke 2002) Development Application of the Globus Toolkit The prime driving force behind the use of the Toolkit is focused on scientific and business applications, extremely high end applications. Examples of such application driven projects include this simulation from computational scientists at Brown University to simulate arterial blood flow. Another example of how suitable this technology is for simulation experiments is with the Southern California Earthquake Centre that simulated not only the earthquake itself but also calculating the effect of the shock waves. The Globus site itself is a community that is dedicated to the development of Grid systems and applications like shown above and are vital to many industries that are part of the Engineering and Physical Sciences expertise and provides high level abilities with respect to testing measures and rendering capabilities required by those in this field. .NET and CORBA Disciplines The question to be asked here is if the .NET and CORBA technologies would be capable of performing this type of peer-to-peer collaborative abilities. With the recent release of the new peer-to-peer technologies by Microsoft and their .NET platform, it has been found that those systems running .NET inside the new server configuration and Windows server technologies, the speed of rendering 3D animated applications could exceed all expectations. With respect to CORBA, Globus’ standard statement reflects that “the Common Object Request Broker Architecture (CORBA) defines a standard Interface Definition Language (IDL) for inter-language interoperability, a remote procedure call service, and a variety of more specialized services such as a trader (for resource location). Like Java, CORBA provides important software engineering advantages, but it doesn’t directly address the challenges that arise in Grid environments such as specialized devices and the high performance required by many Grid applications. Again, Grid and CORBA technologies are complementary, not competing. We are working with several groups to develop CORBA/Globus interfaces.” (Globus n.d.) IBM’s influence with Grid Technologies In 2002 it was announced that “it had become the first major IT vendor to provide commercial support and distribution of the Globus Toolkit” (Shread 2002) which would then create support for grid computing for an “IBM eServer p690 supercomputer sold to the Research Councils UK that would form one of the major processing engines of the UK National Grid” (Shread 2002). This operability between a powerful server computer and other supercomputers that would form part of web of supercomputers would in fact “offer the potential to double the processing power of the UK Grid and be used in everything from drug design for commercial use to studying the structure of the earth” (Shread 2002) thus making the ability for any user or corporation far more accessible due to its advantageous platform compatability. Since IBM is offering the toolbox free of charge and is middleware, the ability to work with “third party software technologies such as Avaki, DataSynapse, Entropia, Platform Computing and United Devices” (Shread 2002) an enviable open source technology. Conclusion The Globus toolkit is not a competing software package for others like .NET and CORBA, but, rather a distinctive and layered software system that, in all appearances, runs seamlessly inside your application sharing system. It is also evident that this is a software package that is engineered for sophisticated technology application that unless the server technology within your own computing environment were extremely fast, this is an alternative method of peer-to-peer network sharing for such rigidly memory intensive applications that engineers and physicists require the ability to effectively render their research quickly and efficiently. Works Cited Foster, C. Kesselman, S. Tuecke. 2001, The Anatomy of the Grid: Enabling Scalable Virtual Organizations. International J. Supercomputer Applications, 15(3), [online] Retrieved 16 Mar 2006. Available at: http://www.globus.org/research/papers/anatomy.pdf. Foster, C. Kesselman, S. Tuecke. 2002. The Physiology of the Grid: An Open Grid Services Architecture for Distributed Systems Integration. [online] Retrieved 16 Mar 2006. Available at: http://www.globus.org/research/papers/ogsa.pdf. Globus. About the Globus Toolkit. [online] Retrieved 16 Mar 2006. Available at: http://www.globus.org/toolkit/about.html Globus. Frequently Asked Questions. Doesn’t CORBA solve these problems? Globus.org. Globus Toolkit History. [online] Retrieved 16 Mar 2006. Available at: http://www.globus.org/toolkit/presentations/GThistory2.ppt@261,1,GlobusToolkitHistory Lumb, Ian. 2003. Dissecting the Hype Around OGSA and Grid Standards. [online] Retrieved 16 Mar 2006. Available at: http://www.gridcomputingplanet.com/opinions/article.php/1569321 Shread, Paul. 2002. Globus Gains Commercial Acceptance. [online] Retrieved 16 Mar 2006. Available at: http://www.internetnews.com/infra/article.php/1431371 Read More