What if we were “Home schooled”?
“Home schooled?? What? Which crazy parent would put their child at home and presume to give all the exposure he needs within those walls? What about bragging? “My kid is studying in so and so fancy pants school. What about yours?” Ok. Leave all that. Let’s talk about the kid. Did you think about Socializing? What about some friends for him? What about getting a life? What about this! What about that?”
I know, I get it. Home Schooling is not even a thing in India. Our parents and the society we live in, believe in public education at a level that even film makers who go out of scripts don’t care to make a movie on home schooling, let aside practicing it.
Coming from India, we do and follow many activities as a tradition, but sometimes when we look at it in a scientific way, we come to know the reason behind us doing certain things the way our ancestors did and why derailing away from that will not be the best thing to do.
Also, in contrast, things which we follow as a tradition may not be applicable in the 21st century we live in too. Regarding that, I remember a saying from my mom when I was a child, “Don’t cut your nails, after 6 PM. Its unpropitious to our family members”. As an obeying son, I made sure I never did that, hoping it would not bring bad luck to my family members. But as I grew up, I tried to explore the scientific reason behind it, and realized that during my ancestor’s time, probably a century back when it originated, they never had access to “Electricity” or any sort of light source after sunset, which would make them physically vulnerable to cuts while clipping their nails in the dark or even wound the family members if they accidentally stepped on those clipped nails. Things like that may be applicable before 18th or 19th century, but I believe it would not be the case in this era.
Likewise, in our IT field, there came certain advancements in the recent past, which demanded a change to our existing IT practices. One such thing, would be on the Analytics side.
We have been collecting data from devices, and developing Machine learning algorithms for decades now, in the venture of building a deep learning program and statistics over the data we accumulated over years. And, AI based companies like DeepMind has made remarkable achievements in the AI area by making robots walk on its own like a human through deep learning algorithms. Those were developed over decades of research and the list goes on and on.
But let’s talk something about the area we (GAVS and its customers) are into, ITOA.
In a layman’s language, the architecture for an ITOA program typically goes like this,
Devices (Sensors, Agents, Gateways etc.,) → Analytics Engine (in Cloud or On-premise infra) → BI dashboard (Graphs, Charts, Cards etc.,)
In this architecture, typically all the data from the devices are sent to the central analytics engine where all the learning and prediction happens, and this engine is made available in the cloud or on-premise infrastructure.
Data scientists usually prefer this approach as they get the raw data from the devices, which would help them predict the incident more accurately. And like they always preach, its “More Data = More accuracy”, and I don’t completely disagree with them.
But like every other approach, there is a downside to this as well. When trillions of telemetry data gets passed to the Data center, it puts immense load on the storage and the analytic engine’s computing power, making the business invest on more Capex, by buying more hardware resources. On the other side, the network traffic will also be high during data transmission, which will hinder the real-time performance SLAs.
We face these issues every now and then, and most of us wish something revolutionary happens in this space. The answer to this is, Edge computing, rated as one of the top strategic technology trends for 2018 under Digital category by technology research giants like Gartner in their report.
With Edge Computing, things are little different, the good news is, you don’t need to dispose any component of your existing architecture, and not worry about adding more Capex for incorporating a new technology. After all, a better ROI is something which we always aim, right?
So, what is Edge computing?
As per definition it is, “A method of optimizing cloud computing systems by performing data processing at the edge of the network, near the source of the data.”
Again, in Layman’s language, its Analytics at device level. So, before the data reaches the cloud or the on-premise data center, certain amount of analytics would already be performed at device level thus reducing the load on the network traffic, storage space and computing power of the central analytics engine.
To explain better, imagine a situation, where trillions of telemetry data is sent every hour to a data center for analysis, which requires 100s of computing resources to process, where 30% of the analytics are performed on repetitive data patterns. Those 30% can be done at device level with less computing power by avoiding network traffic and storage consumption at the central data center, this is how edge computing works.
Now you may have these questions,
“If this was that simple, why was it not done in the past? Why is it booming now?”
“Why should I re-architect my solution/products on something which nobody else does?”
And I think its valid. Reasoning is always good, it allows us to make informed, reasoned decisions and valid inference with structured reasons and evidence without sticking to the traditional practices.
Like the nail story I told you, we need to compare the advancements of the present with the past. In the past the computing power of devices have always been too low. Intel’s 8080 microprocessor had only 2Mhz of processing power, whereas the latest ARM chips used in our mobile phones have a minimum processing power of 2Ghz with 8 cores, which is 2000-5000 times powerful. And with the advancement in technology, our practices should also evolve, don’t you think?
In an article published during late 2016, by an American telecommunication company named Sprint, they have stated that, over implementation of Edge computing in their company, they have seen four major advantages. There were able to get significant benefits in the areas of Real-time or near real-time data analysis, Lower operating costs, Reduced network traffic and Improved application performance.
And not just Sprint, many other global leaders like Nokia, Dell, Cisco etc., have already started implementing edge computing couple of years ago in their products and services.
Speaking of implementation, if you could look closely, Edge computing is not much different from Home schooling, I would say, very similar in terms of implementation.
- In the traditional ITOA architecture, you send the data to the Cloud or On-Premise DC, like you send your kid to a school.
In Edge Computing, you get the data homeschooled by analyzing it at device itself. - In the traditional ITOA architecture, you can’t have real time analysis of data due to network traffic constraints, like you can’t watch your kid every now and then, like you would if he is at home.
In Edge Computing, your data is processed at device level where it originated, so no concept of network traffic here. - In the traditional ITOA architecture, your CapEx and OpEx are usually costlier and its gets higher even now and then, just like how your kid’s school fees get increased every year as he graduates.
In Edge Computing, your data is processed at device level where it originated, so no concept of CapEx and OpEx here. - In the traditional ITOA architecture, you would have take special measures about the security involved at levels, right from data transfer till the final application, like you would have to, while sending your kid to a school. You need to choose the School bus for his transport or drop him at school and pick him up, you need to make sure that the school has CCTV installed, Compound fenced, security personal stationed, perform security checks and clearance procedures for visitors, etc.,
In the Edge Computing, your data gets processed mostly at device level, so no additional security involved than the usual.
Though homeschooling may not be an appropriate solution to your kid, looking at the benefits of Edge computing, it appears to be convincing to get our devices homeschooled.
So, without abiding to tradition, let’s get our “Things” home schooled, shall we?
All images are picked from external sources.
