Cloud storage of data was seen as an environmental savior ten years ago. But extreme increases in data streams are expanding the cloud's carbon footprint, so finding more energy-efficient ways of computing seems more urgent than ever.
Last year, most of the world's leaders came together at COP26 to make new environmental pledges. The goal is to keep the global temperature rise below 1.5 °C, and pressure is growing on businesses to step up and play their part.
When we moved into the internet era 25 years ago, most people only saw the environmental benefits, how the internet and email cut down tonnes of physical resources like paper. Ten years later, cloud storage seemed to be the new ecological savior in the IT world. It reduced energy consumption for companies compared to having on-site servers, reduced greenhouse gas emissions, and contributed to dematerialization. But quite soon, we realized that the carbon emissions produced by manufacturing, powering, and cooling computers, smartphones, and the enormous data centers where cloud data is stored even out these benefits.
Today Information and Communication Technologies (ICT) significantly contribute to carbon emissions. ICT, mostly made up of data centers and cloud computing accounts for between 4-6% of the world's electricity consumption.
When organizations examine the environmental impact of their IT operations and look for ways to lower their carbon footprint, they still often move their data to the Cloud, reducing emissions from their digital applications and infrastructure. But how much environmental impact does this have when the reality is that you are just moving that carbon footprint along the supply chain rather than removing it entirely?
Let's take an easy example. Say you are saving a standard business document. If you save it to your hard disk, it requires about 0.000005 kWh per gigabyte to save your data. On the other hand, if you save it on the cloud, a Carnegie Mellon University study concluded that the energy cost of data transfer and storage is about 7 kWh per gigabyte. This is because cloud data is not stored in actual clouds but in buildings — massive structures filled with thousands of hard drive-bearing racks using tremendous energy. There are millions of data centers worldwide, some taking up nearly 200 acres of land apiece.
It takes energy to get the data to the data center — miles of fiber optic cables studded with other fixtures of internet infrastructure that all require power along the way. At the center, your data is stored multiple times on hard disks. The constant activity of all those disks creates a lot of heat, necessitating energy-intensive air conditioners to protect the equipment from overheating.
Saving a document to the Cloud is convenient since it takes it out of your hands. You don't have to worry about losing it and can access it anywhere. It also gives you the peace of mind of a secure backup, but at what cost?
All the big players in the industry have invested a lot of money in new Clouds. For example, Microsoft is now changing their private Office customers' possibilities for saving data to their hard drives. With Windows 11 standard version, you get tied up to Microsoft 365, which automatically saves all your data in the cloud. You don't even have the option to keep it on the hard drive in the less expensive versions of Microsoft 365.
So even though Microsoft has high rising aims of being carbon neutral by 2030. Right now, they are escalating emissions in high numbers by transferring an enormous stack of private customers, with no real need for cloud storage, into their Azure cloud. And the reason for this isn't hard to figure out. Cloud storing might be accessible for these new transferring users for six months, maybe a year, but of course, it's a giant cash cow, in the long run, set up as it is a subscription service, at the moment with a fee of 59.99 $/year for one person and one terabyte.
And it's easy for Microsoft to keep customers once they're aboard. If you have already stored all your data in the cloud, you have to go on being connected with Microsoft if you want to keep it. It is written out in the information to Microsofts private customers that you, as customers, are obliged to be connected to the Azure cloud at least once a month. If not, they'll risk slower connection speeds and software failures.
But are there any sustainable alternatives to massive cloud storage, then? Of course, there is.
One company that recently has developed a sustainable alternative to dependence on the Cloud is IoE Corp, the fast-progressing R&D Start Up I'm working for myself now. Although, IoE Corp's target is not private customers who need to save random documents. They provide a solution for the new world of smart cities, for incorporated AI services which produce vast amounts of data, a world of autonomous automation, and massive IoT and IIoT systems.
Since the meaning of smart cities has started to be more about silly apps than infrastructure solutions, IoE Corp doesn't use that word. IoE Corp instead talks about developing informed infrastructure, and their tech solution has wide scalability. It might be deployed to a system of traffic lights in an area, one smart home, or a whole integrated city of different AI solutions.
What IoE Corp does is that they stay away from the WWW. They work the internet core, putting up a secure blockchain for systems of nodes that all control each other, a decentralized system where the data is spread out, thereby not requiring large server halls where machines need to be cooled down. The idea is to create sustainable, secure systems beyond the cloud without cloud storage.
– Some services today require a cloud connection, but when it comes to IoT, then for the sake of security, stability, and privacy, avoid the cloud as far as possible. Web service providers built the cloud to optimize the web. Mattias Bergstöm, serial entrepreneur, futurist, and CTO of IoE Corp, explains that massive IoT and the web are not a good fit.
The technology he has created is based on a human-first architecture and runs on a truly decentralized infrastructure. It's named Eden, and the Eden system is a decentralized, autonomous, portable, secure virtual infrastructure for managing clustered workloads over depos (decentralized pods) and services that facilitate declarative configuration and automation.
The decentralized model is based on scalable device clustering, where adding new devices as nodes is easy. This makes it possible for any device to contribute computing resources over an intelligent mesh network so that computing can happen where it is needed and close to where it will be used. Eden is developed via quantum-safe tunnels, using polymorphic encryption keys and a blockchain with consensus to verify the data moved between the nodes over the tunnels, thus creating trusted data-walled gardens.
The orchestration of computing and storage is done via service manifests that describe service rules, policies, and logic. An autonomous knowledge-based AI manages the underlying orchestration mechanics using network consensus over the blockchain as a deciding mechanism. At last, the cluster topography is dynamically updated by the orchestration to fit the current workload. Eden system service depos are generated and deployed similarly to container images; the depos are MPI (Message Passing Interface) cluster enabled from the start.
Implementing these innovations into IoT devices deployment aids in keeping the energy and cost levels of computing at a sustainable level. Adding a viable option to the data centers that will constantly have to be readjusted to maintain a reliable and safe operating system that complies with sustainable computing requirements.
Another aspect of sustainable computing that most companies neglect is that the language the code is written in makes a tremendous difference in how much energy is used when running a service or program. We must be conscious of the wasteful nature of some programming languages and try to implement greener options if we truly want a world of sustainable computing to become real.
There is a big difference in what kind of programming language you choose. Compiled languages like C, C++, Rust, and Ada rank as some of the most energy-efficient, while interpreted languages like Perl, Python, and Ruby are among the least energy efficient. On average, if sorted by their programming paradigm, the imperative languages needed the least amount of memory, followed by the object-oriented, the functional, and finally, the scripting languages.
One of the most wasteful interpreted languages is Python. It's extremely popular. Global apps like Youtube and Netflix are, by and large, programmed in Python. It is also often used when teaching programming due to its educational qualities and ease of use.
But of course, all this has to end soon if our goal is to implement sustainable computing. Therefore it was easy for IoE Corp, a company that works with edge commuting and sustainability at the core, to choose a compiled language to program in instead of C or C++ Rust is used.
– The idea of sustainability through design is fantastic, but we should remember that most of the wasteful behavior comes from implementation, interpreting programming languages that are wasteful, and they use the Cloud; that bottom line depends on wasteful code, Mattias concludes.
There is a tremendous and urgent need to start living sustainably. We all have to do our best — governments, companies, and ordinary people — to keep the global temperature rise down as much as possible. And when it comes to sustainable computing, IoE Corp's way forward is to be sustainable for real: by using edge computing, a compiled programming language, and going beyond the cloud.