When we are given a problem to solve in IT (software or hardware), we have a few things in our laundry list to align for the the developed system. To name a few: Scalable, Extendable, Reliable, Secure, Fault tolerance, avoid Single Point of Failure (SPoF) etc. They're definitely absolutely required. But I feel that we're narrowing down too much on getting a good system that we most of the time miss the big picture. We try to build system that's unnatural within this natural universe we live in.
For decades, we have accepted the notion that whatever we build in IT are artificial and meant to remain artificial which is true but no one is stopping us to get influenced by nature - in fact - nature is the humungous system that survives through all these millions of years proving that it is reliable, scalable, fault tolerant, very few SPoF, you name it. Actually other science and technology disciplines have always adopted the notion to get influenced by "nature" the but in IT it seems to have very negligible influence.
It's time to think and see with a brand new way when we design a IT system - think to keep the design organic. One very important factors of design that's abandoned in nature is the "Distributed Processing". To be more specific, nature has created this world in the essence of distribution in processing and in risks while keeping a centrally managed control and monitoring system. Apache Hadoop is one of the system that became tremendously successful by keeping the distributed processing and distributed risks at the core of it's design while keeping a centrally managed control point. In Hadoop the Data node (highly distributed in processing and risks) and the Name node (central management) are example of that.
Another design philosophy of natural design is the Multilevel Security. Think about when we eat something or as simple as breathing, the first level securities are always in place at the interfaces thought they're not very strong and definitely don't intended to replace the real antibody system in side.
Below checklist can be run on a software or IT systems to determine if its design is organic:
1. The system doesn't consume or produce more data than it absolutely needs. There should be a quality ratio on how much data a system uses and how much data it absolutely needed to perform that process. Data derivation should be a must.
2. It recycles its own garbage data i.e. the recycling should be part of the design of the systems. When we design a system we mostly design to produce data but rarely design on recycling the information that is used by it. For example, when a program or process is no longer in need of any specific data then there has to be a way to recycle that information.
3. It collaborates to perform to its best i.e. it performs better or can do more when it's able to collaborate with other systems.
4. The multilevel security is in place. Traditional password based security isn't unnatural as when you break it once you're in full control. The natural security would like at the entrance of the system it would have a first pass of security (biometric authentication would be first choice) and within the system, every movement is monitored and an army of security process (just like our anti body, RBC, WBC etc.) would be ready to stop them while any suspicious move is taken. For that pattern recognition, profiling of user's activity etc. would be used to differentiate between good move and bad move.
5. The system should be intelligent in nature. This feature is required to build most of the above mentioned features but it should have more than just security, data recycle etc. The intelligent system may or may not be Artificial Intelligent (AI) systems but it ought to be able to intelligent enough to interact with environment like network quality, data quality, availability of dependent system and based on those it can upgrade or downgrade its features availability.
I've been thinking about this topic for quite some time so thought to start putting it to my blog if it can be taken any further. Also I would like go bit in depth for some of the check list items in future posts.
For decades, we have accepted the notion that whatever we build in IT are artificial and meant to remain artificial which is true but no one is stopping us to get influenced by nature - in fact - nature is the humungous system that survives through all these millions of years proving that it is reliable, scalable, fault tolerant, very few SPoF, you name it. Actually other science and technology disciplines have always adopted the notion to get influenced by "nature" the but in IT it seems to have very negligible influence.
It's time to think and see with a brand new way when we design a IT system - think to keep the design organic. One very important factors of design that's abandoned in nature is the "Distributed Processing". To be more specific, nature has created this world in the essence of distribution in processing and in risks while keeping a centrally managed control and monitoring system. Apache Hadoop is one of the system that became tremendously successful by keeping the distributed processing and distributed risks at the core of it's design while keeping a centrally managed control point. In Hadoop the Data node (highly distributed in processing and risks) and the Name node (central management) are example of that.
Another design philosophy of natural design is the Multilevel Security. Think about when we eat something or as simple as breathing, the first level securities are always in place at the interfaces thought they're not very strong and definitely don't intended to replace the real antibody system in side.
Below checklist can be run on a software or IT systems to determine if its design is organic:
1. The system doesn't consume or produce more data than it absolutely needs. There should be a quality ratio on how much data a system uses and how much data it absolutely needed to perform that process. Data derivation should be a must.
2. It recycles its own garbage data i.e. the recycling should be part of the design of the systems. When we design a system we mostly design to produce data but rarely design on recycling the information that is used by it. For example, when a program or process is no longer in need of any specific data then there has to be a way to recycle that information.
3. It collaborates to perform to its best i.e. it performs better or can do more when it's able to collaborate with other systems.
4. The multilevel security is in place. Traditional password based security isn't unnatural as when you break it once you're in full control. The natural security would like at the entrance of the system it would have a first pass of security (biometric authentication would be first choice) and within the system, every movement is monitored and an army of security process (just like our anti body, RBC, WBC etc.) would be ready to stop them while any suspicious move is taken. For that pattern recognition, profiling of user's activity etc. would be used to differentiate between good move and bad move.
5. The system should be intelligent in nature. This feature is required to build most of the above mentioned features but it should have more than just security, data recycle etc. The intelligent system may or may not be Artificial Intelligent (AI) systems but it ought to be able to intelligent enough to interact with environment like network quality, data quality, availability of dependent system and based on those it can upgrade or downgrade its features availability.
I've been thinking about this topic for quite some time so thought to start putting it to my blog if it can be taken any further. Also I would like go bit in depth for some of the check list items in future posts.
No comments:
Post a Comment