35
loading...
Dataflow Programming & it's impact on future
Dataflow programming (DFP) introduces a new programming paradigm that internally represents applications as a directed graph, similarly to a dataflow diagram. Applications are represented as a set of nodes (also called blocks) with input and/or output ports in them.
These nodes can either be sources, sinks or processing blocks to the information flowing in the system. Nodes are connected by directed edges that define the flow of information between them.
Due to social media (2000s) and IoT (2010s), there’s a need to analyze data quickly and derive timely insights. There’s also a need to explore and exploit parallel architectures. Dataflow programming aids parallelization without the added complexity that comes with traditional programming.
Dataflow programming includes two sub-fields: Flow-Based Programming (FBP) and Reactive Programming. Dataflow programs can be created in text or visually. There are many languages that support this paradigm.
Benefits
The key benefit of Dataflow programming is that, in dataflow, more than one instruction can be executed at once. Thus, if several instructions be- come fire able at the same time, they can be executed in parallel. This simple principle provides the potential for massive parallel execution at the instruction level.
The key benefit of Dataflow programming is that, in dataflow, more than one instruction can be executed at once. Thus, if several instructions be- come fire able at the same time, they can be executed in parallel. This simple principle provides the potential for massive parallel execution at the instruction level.
DFP has enabled many visual programming languages that provide a more user-friendly interface so that even non-technical users can write programs. Such languages are also suited for rapid prototyping.
DFP promotes loose coupling between software components that’s well-suited for service-oriented architectures. Its functional style makes it easier to reason about system behavior. Its simplicity is in its elegance.
DFP is deterministic. This enables mathematical analysis and proofs.
Main Characteristics
Dataflow programming emphasizes the movement of data and models programs as a series of connections. Explicitly defined inputs and outputs connect operations, which function like black boxes. An operation runs as soon as all of its inputs become valid.
Dataflow programming emphasizes the movement of data and models programs as a series of connections. Explicitly defined inputs and outputs connect operations, which function like black boxes. An operation runs as soon as all of its inputs become valid.
Some other alike Programming Paradigms
Reactive Programming is a declarative programming paradigm concerned with data streams and the propagation of change. With this paradigm, it’s possible to express static (e.g., arrays) or dynamic (e.g., event emitters) data streams with ease, and also communicate that an inferred dependency within the associated execution model exists, which facilitates the automatic propagation of the changed data flow.
Reactive Programming is a declarative programming paradigm concerned with data streams and the propagation of change. With this paradigm, it’s possible to express static (e.g., arrays) or dynamic (e.g., event emitters) data streams with ease, and also communicate that an inferred dependency within the associated execution model exists, which facilitates the automatic propagation of the changed data flow.
Flow-based programming (FBP) is a programming paradigm that defines applications as networks of “black box” processes, which exchange data across predefined connections by message passing, where the connections are specified externally to the processes. These black box processes can be reconnected endlessly to form different applications without having to be changed internally. FBP is thus naturally component-oriented.
FBP is a particular form of dataflow programming based on bounded buffers, information packets with defined lifetimes, named ports, and separate definition of connections.
Impacts of Dataflow Programming in the future
I believe that several factors are motivating a renewed interest in the design and implementation of scalable dataflow processors. These include the recent technological advances in increased chip density; the complex interconnections among multiple processing elements on a single chip (Network-on-a-chip); the hardware complexity of the superscalar, super pipeline, and VLIW architectures and the diminishing performance gains of these systems with additional hardware; the large and multi-level caches; the compliers that perform extensive global and intraprocedural analyses to extract as much parallelism as possible, and finally, the success of the recent dataflow projects (Recent Dataflow Projects section).
I believe that several factors are motivating a renewed interest in the design and implementation of scalable dataflow processors. These include the recent technological advances in increased chip density; the complex interconnections among multiple processing elements on a single chip (Network-on-a-chip); the hardware complexity of the superscalar, super pipeline, and VLIW architectures and the diminishing performance gains of these systems with additional hardware; the large and multi-level caches; the compliers that perform extensive global and intraprocedural analyses to extract as much parallelism as possible, and finally, the success of the recent dataflow projects (Recent Dataflow Projects section).
At the same time, if dataflow architecture is to address the challenges of future processing architectures containing hundreds, if not thousands, of processing elements, it is necessary to evaluate carefully different forms of dataflow organizations for their suitability for implementation.
35