Replacing Servers with Smart Devices and IoT
The world is now moving towards alternatives of products that are more ecological. Smart devices nowadays possess processing power more than that of supercomputers in the earlier days. Smartphones, specially, are more powerful than the computers used by NASA on the Apollo mission. Devices like smart TVs, smart refrigerators and other IoT devices have the potential to replace the server farms. This can be done by creating clusters of these devices and using small amount of space on each device to store and process data. This will create an opportunity for consumers to rent out the storage not needed by them and provide processing power and storage to those in need. This technology has the potential to reduce the number of servers on the face of our planet. Also, this is more ecological than servers as these devices used a small fraction of power than that required by servers.
There are more than 80 million servers in the world powering the internet. These servers consume a lot of energy and are becoming a major cause of global warming. The idea of this paper is to develop a technology in which the storage and processing power of smart IoT devices can be used to store and process data of the population and reduce the number of server farms. This can be done by creating a platform that let users rent out their digital storage to the ones in need. Local clusters of devices can be created that distribute the workload and work efficiently consuming less power. This will bring down the number of servers and also will be a big step towards ecological technology.
Since large-scale data processing relies on distributed computing, by reducing individual task size in a computing job we can run these tasks using devices with limited memory, such as smartphones and smart TVs. The smart devices like smartphones, smart TVs and smart refrigerators can be interconnected through the IoT. Small amounts of space can be rented out by the user and this network of devices would possess the ability to process and store data for people. These clusters of devices can be made to work synchronously to improve the efficiency of the process. The data should be stored in a compressed form to minimize the time lag. This way a cluster of smart devices can work in the same way as a server but consume much less power and be both more ecological and economical.
Ecological Footprint of Servers
An average server can consume between 500 – 1200 watts per hour depending on the use. If the average use is 850 watts per hour, multiplied by 24 that equals 20,400 watts daily, or 20.4 kilowatts (kWh). Multiply that by 365 days a year for 7,446 kWh per year. This is more than the consumption of an average American household for a year. IT-related services now account for 2% of all global carbon emissions, according to a new Greenpeace report which is equal to the aviation sector. Not only emissions, there are many other issues like e-waste and coolants which cause a lot of problem.
Comparison of Consumptions and Processing Power
The power required by smart devices is very less when compared to the servers. An average smartphone has a 3.5V 3000 mAh battery which means it’s power consumption leads to 5.45 Watts per hour. Assuming that it needs to be charged twice per day and takes 2 hours to full charge the power consumption per day is 21.8 Wh which multiplied by 365 gives approx. 8000 Wh or 8 kWh. This is quite less when compared to 7,446 kWh per year and also, we should keep in mind that this power is used not only for the server usage but also for other uses of the consumer.
To compare the processing power of different devices, we use GFLOPS (Floating Point Operations per Second). 1 GFLOPS device will perform 1 billion operations in a second. The processing power of an average server is 500-600 GFLOPS and that of a smartphone is 35 GFLOPS so to equal the power of a server a cluster of 18-20 smartphones (ignoring more powerful devices like the PS4 which has a processing power of 1.4 TFLOPS). A cluster of 20 devices multiplied with the power consumption of 8 kWh gives us annual consumption of 160 kWh so we can see that this alternative technology is quite ecological and economical.
All new technologies come with their own pros and cons. The distributed processing and storage technology is more ecological and economical but there are some disadvantages which make its implementation difficult. It is not easy to convince people to rent their personal devices for processing and storing data of unknown people and at the same time it is very difficult to convince them to store their data on some other device. This problem can only be overcome by building a reputation amongst the people. Another problem is that the processing speed of a cluster cannot match that of a server and this may sometimes result in customer dissatisfaction. To conquer this problem, more and more clusters of devices have to be made available so that the speed becomes better.
Comparison of processing speed – image from https://ubispark.cs.helsinki.fi
Currently this technology might not sound very practical and it is true that it is difficult to implement but in the future when the problem of a more ecological and economical way to replace servers arises, this will be the solution.
There has been an active research on this technology in Finland.
You can check it out here.
About the Author
I’m Sahil Chaudhary, a 17 year old from Kanpur, India.
I have been developing mobile apps and websites since the last three years and also I have been working actively in fields of AI, IoT and Cyber Security. Since the beginning, I had the urge to share whatever knowledge I had with other people and connect to new people so that’s why I started writing blogs. I am currently a Guest Blogger for Nexxy Technologies Blog. you can shoot me a direct mail HERE or follow me on Instagram