Not all the problems that computing works with can be solved with a single very powerful processor, sometimes it is necessary to use a set of processors to solve the problem, but the cores of a single computer are not always enough and it is necessary to create extremely complex infrastructures composed of several computers, which we often cannot contain in one place and this is where both cloud computing and distributed computing emerge.
Cloud computing refers to the use of high power remote computers to run programs, but it is not the only way a computer can be used remotely, with distributed computing being one of those ways. to use computer resources. remote computing, but its operation is reversed.
What is distributed computing?
Distributed computing refers to a way of executing a program in which a group of processing units, geometrically dispersed but connected across the same network, each participate in its resolution. These processing units do not necessarily have to have the same power in common and there may be units with different characteristics.
Standard distributed computing uses a centralized unit, which is responsible for distributing the work to be done between the different processing units available to the rest of the processing units in the network, because they do not have good access to the resource. Instead, they act passively, waiting for the central unit, which has access to these resources, to send them the instructions and the data to be processed. Once a processing unit has completed the assigned work, it sends the result back to the central unit which is the one that manages the resources and waits for new tasks to be assigned to them.
There can be two ways to manage the division of labor, one of them is to use a general-purpose processor running a program, the other is to use a dedicated unit which keeps track of all the processing units that are active at all times. and which are not.
Distributed mesh or grid calculation
One type of distributed computing is mesh computing, but the main difference is that in distributed computing, an element must act as a conductor, which is in charge of resource management and makes all processing units act at the same time. . in unison by dividing a common problem which has been divided into several parts.
In the case of mesh computing, each of the processing units has its own resource manager and a centralized resource management system is not used. It is widely used to solve problems which not only can be divided and solved independently, but which does not have great dependence and iteration between different parts, so that each processing unit can work completely independently.
At present, they are the most widely used type of distributed computing, especially in the world of scientific computing, because the problem of having a conductor responsible for managing all the resources at the same time is avoid.