You might be familiar with the works of a power grid, where various sources of electricity are linked together to supply power to a certain geographical location and the consumers who live there. Reflecting on your day-to-day and your monthly electricity bill, it is evident that power usage varies daily for reasons like the weather, or activity level. Due to this variation in demand, the load on the power station is inconsistent and constantly changes. The power station must therefore be able to scale up or down when demanded to meet the requirements of consumers. A power grid can meet these demands, generating power from an array of different sources and efficiently distributing the load amongst its connections. Whereas a grid-less power supply runs the risk overreliance on a single source, which can lead to price hikes during high demand periods and even outages should something fail.
What is Grid Architecture?
Grid architecture is a similar concept to an interconnected power grid in that computers are linked together to provide a greater resource. Grid architecture uses load sharing software that distributes the load across many smaller components in is an arrangement of computers, connected by a network, where unused processing power on all the machines is harnessed to execute tasks faster. Tasks are distributed among the machines thereby minimizing wasted resources and increasing efficiency. Grid architecture is commonly used for complex computational processes such as protein folding, financial modeling, earthquake simulation, and maybe even just to maintain daily performance levels of your enterprise archive. Some of the advantages of grid architecture include:
- Lower costs. No need to buy one large, expensive server to complete a job that can be split up and executed by several small, inexpensive servers. Upon job completion, the results from each machine can be amalgamated and analyzed as a whole.
- Increased efficiency. Jobs can be reallocated to idle servers. Policies can be implemented that allow jobs to only go to servers that are lightly loaded or have the appropriate amount of memory/CPU characteristics for the particular task.
- No single point of failure. If one of the servers within the grid fails, other resources within the system are made available to pick up the slack.
- Scalable. If you require additional resources, just plug them into the system by installing grid client on additional servers or desktops. These additional resources can also be removed just as easily.
- Automatic management. A client resides on each server to send information back to the master server, telling it what type of availability or resources it has to complete incoming jobs. The software coordinates the entire process without the need for manual execution.
Grid architecture, as with a power grid, links resources together in a way that allows one source to access and leverage the power of all the various different sources in the system. The cluster used in grid architecture can vary in size, from a small desktop to several networks, making it versatile and customizable depending on the user’s needs, project size and available resources, and allowing full maximization of all available resources.