Smart Computational Grids: Infrastructure for Tackling Global Integration Challenges
News Release from:
15 August 2012
Vladimir Getov, Senior IEEE member and professor of distributed and high-performance computing at the University of Westminster, discusses smart computational grids in the wider context of a smarter planet vision. Vladimir highlights that being connected is not sufficent to address global challenges and that we need to address the need for making global systems better.
The main challenges for our planet are becoming grimly clear – the first decade of the twenty-first century has been a series of wake-up calls, with a single subject of focus, the reality of global integration. The most notable ones include:
• Climate change and global warming
• Population growth
• Frozen credit markets and limited access to capital
• Energy crisis including energy shortfalls and erratic commodity prices
• Healthcare management and delivery around the world
• Increasingly complex supply chains and empowered consumers
Just being connected is not sufficient to address our challenges. There is a need to make these global systems better. Energy systems – 170 billion kilowatthours wasted each year by consumers due to insufficient power usage information. Healthcare systems – that don’t link from diagnosis, to drug discovery, to healthcare deliverers, to insurers, to employers while facing pandemic challenges such as the outbreak of swine flu. Traffic systems – congested roads cost us billions of lost hours and billions of litres of wasted petrol as well as having a huge impact on the air quality.
The smarter planet vision which was initially based on the launch of IBM’s campaign in 2008 addresses the above and similar challenges by introducing three key elements:
•First, our world is becoming instrumented. Sensors are being embedded across entire ecosystems, supply-chains, healthcare networks, cities and even natural systems like rivers
•Second, our world is becoming interconnected. Systems and objects can now “speak” to one another. Soon there will be a trillion connected and intelligent things – cars, appliances, cameras, roadways, pipelines, pharmaceuticals, and even livestock. The amount of information produced by the interaction of all those things will be unprecedented
•Third, all things are becoming intelligent. Advanced analytics can turn the mountains of data from these systems and objects into decisions and actions that make the world smarter
The agenda for a smarter planet is a transformational one to create and manage a new future for these instrumented, interconnected, and intelligent systems that come together and to bring solutions that chart the course for real-time collaborative ecosystem management. In order to start tackling the above challenges we need qualitatively new large-scale computing infrastructures with smart properties.
In recent years the concepts and implementations of modern distributed computing infrastructures have been developing rapidly. Following the initial meta-computing ideas from the 90s, the professional community has been advancing and developing grid computing systems. A computational grid is a federation of computer resources usually coming from different administrative domains but put together for the solution of a specific large-scale application. Among other specific features grids are characterised by the introduction of the middleware layer in their architecture. The middleware is responsible for the distribution and coordination of program execution among the participating resources in the grid infrastructure. Traditional grid systems however are not “smart”. On the contrary, they require significant human component in their operations including deep understanding of their architecture.
As a next step in the development of novel distributed systems, the smart computational grid paradigm emerged several years ago after the introduction of the “invisible grid” concepts. In smart computational grids, providing support for a variety of autonomic properties is of primary importance in the construction of these high quality and large-scale complex systems. Indeed, a large proportion of research and development efforts has been invested in investigating the services design methodology in dynamically reconfigurable distributed platforms supporting flexible and fault-tolerant composition and execution of workflows. This has resulted in the introduction of new approaches, methodologies, tools and environments contributing directly to the long-term objectives for developing smart computational grids. Also, the increased interest and motivation in the partnership related to services and service oriented architectures, has proved to be particularly interesting and productive.
The smart computational grids are characterised by a number of novel smart properties, and in particular those related to:
•Advanced grid programming models and workflow management systems, including international efforts such as those related to the SCA initiative by IBM and the “invisible grid” concepts
•Autonomic management of non-functional features in distributed programming
•Complex (web) service orchestration and choreography, such as those related to WSO with BLEP4WS or WSCI
•Modern software engineering approaches such as algorithmic skeletons, design patterns, abstract software component models and related methodologies
•Utility computing achievements, as well as theoretical and practical results from the cloud computing community
Thus, smart computational grids are fast becoming a primary choice for building global infrastructures to address and solve the challenges associated with a smarter planet.