DIET: The Grid & Cloud Middleware

DIET stands for Dual Intent and Entity Transformer and can be used for multiple problem-solving purposes on a quick and efficient basis in the Cloud environment. It is free and open-source software for grid computing. It acts as the middleware that provides services upon getting requests to software applications.

Cloud

• It aims at being dynamically scalable and offering virtualized resources as a service over the Internet.
• It is balanced when it comes to computational power and storage.

Grid

• The grid has high computational power which is a big boon in today’s world. 
• It comes with a drawback though although having high computational capabilities, it’s not easily scalable.

To get the best of both cloud and grid computing, a Grid middleware can be used to manage and harness raw Cloud computational resources.

DIET Architecture

DIET implements service calls, scalable scheduling, and data management which in turn helps to implement a Cloud middleware using it. In this architecture, the client’s job is to transfer the problems to the server, server’s job is to solve the problems. DIET has features like

• Database to store information about the computer’s resources.
• A scheduler that schedules and allocates suitable servers for a given problem using information from the database.
• Monitors in DIET that retrieve information related to computing.

DIET Components

• Client:- An application that uses DIET to solve computer-related problems
• Master Agent (MA):- It takes computational requests from clients and also collects computation abilities from the servers. It chooses the server with the best computation ability available for a given problem. A reference copy of the chosen server is sent to the client. A client can be connected to an MA by a specifically named server or through a web page storing the various MA locations.
• Local Agent (LA):- Its job is to act as a middleman between MAs and servers. The information stored on an LA is the list of requests and the number of servers that can solve a given problem and information about the data.
• Server Daemon (SeD):- A SeD is the entry point of a computational server. It has various functions like maintaining the processor. It stores information regarding the list of the data available on a server and the list of problems that can be solved on it. It also keeps the information about CPU capacity, available memory, and some other data related to the load of the CPU.

Advantages of DIET

• Large and complex problems can be solved using DIET middleware.
• There are many processes to port applications to the middleware such as message-passing, batch processing, web portals, and Grid-RPC systems

Disadvantages of DIET

• Most of current applications are numerical and use libraries like BLAS, LAPACK etc. The integration of such libraries in high level applications using languages like Fortran or C is difficult.
• The computational power and memory needs of grid applications are not available on every workstation.