Hello everyone! As promised, we continue to consider possible applications (services) for the adaptive GONT consensus using alternative gas. In today’s article we will focus on the use of HW ALGas. We will also talk about an example of Apple’s demand and tell you how to put the new “iron gas” model into practice.
HW AlGas – Iron Alternative Gas
To implement new types of video accelerators in GONT we should introduce the concept of HW AlGas.
Let’s consider the Value Chain inside the video processing chip.
This chain reflects the level of interaction of individual IP.
Value Chain: images from a compressed video stream
This real Value Chain decrypts the video. In this case, both SW and HW models of this chain can be represented.
A chain of IP blocks inside a large computer (a video decoder as an example) can be viewed as an AlGas chain. In this chain new algorithms can be built in.
For example, the algorithm of the ad filter hardware (By the way, Apple has such a patent! We will talk about this example a little bit later.)
Dividing HW AlGas into clock domains inside the chip:
The “divisibility by AlGas quarks” philosophy can be different within a chip. The only question is where to put gas meters.
Division of HW AlGas by Clock-domains
HW AlGas routes inside the scale chip:
Example of demand from Apple Corporation
How to implement a new model with HW AlGas (“iron gas”) in practice?
Apple Corporation example:
Adding any filter to the usual flow (in fact, GONT Value Chain) video processing is very ontological (potentially many microservices).
Adding microservice (FILTER) to the usual service
For example, if the task is to implement a TV advertising hardware filter in STB cable TV set-top boxes.
Motivation from real life (from Apple):
FILTER 1 (Microservice 1):
Apple has patented a device for hiding (ignoring) advertising.
FILTER 2 (Microservice 2):
Apple has patented technology that will not allow shooting concerts on the iPhone and iPad.
– CAN GONT easily implement such schemes?
These are all examples for independent implementation on the GONT platform!
You need to implement Set-Top-Box with advanced functionality. At the same time, the advertising filter requires constant “training” on new advertising. And in such a configuration it is convenient to introduce AlGas. Since we have both global service (STB) and microservices (ad filters, filter training, etc.).
There are two main tasks:
- Embed a new algorithm in the video decoding conveyor (through the AlGas mechanism).
- Consider the possibility of parallelization of video processing according to the “one algorithm – many data” scheme. To implement sharding of Value Chain chains.
Thank you for attention! Have a productive day!