Decentralized Computing: More than one is better
June 10th, 2018 — There are many blockchain decentralized computing projects in the market of late, such as iExec, Sonm, Golem, Dadi, Lagoon, and Hypernet that appear to be addressing similar markets. At DeepCloud AI we welcome competition, as a healthy amount of competition is always good. In order for constant progress in technology, companies need to have both a sustainable business model as well as healthy incentives and motivation to work even harder for the betterment of our future. As pioneers in both the blockchain and decentralized cloud computing spaces, we believe that not only is there room for more than one, but that the market is extremely young, and technologies have yet to be implemented in a way that achieves maximum impact. In fact, when you dive into the details, each of the decentralized cloud projects have very different internal goals, and even the target markets they are addressing can be extremely varied or niche.
Let’s take an example and look at Hypernet, which has surfaced many times in our community recently — mentioned as competitors to our product. They are targeting the High Performance Compute (HPC) market, traditionally used for scientific computations, such as physics simulations, weather prediction, and earthquake seismic activity analysis. These computations are very data and compute intensive, requiring very high-speed networks for sharing the data across large clusters such as the Cray Supercomputers’ Titan using specialized hardware like InfiniBand network switches with dedicated high-speed connections across the cluster.
Hypernet believes that their Distributed Average Consensus model will make it easier for such applications to converge on a shared peer to peer network with very high latencies and no predictability of resources at any time. This is a highly ambitious goal; researchers for many years have been working to solve this problem, without much success so far.
HPC applications fall under a very narrow and specific market segment of the scientific community and require special tools and engineering to make them work even in traditional environments. Complex models like weather simulations can often take days, weeks or more than a month to converge even on the largest Supercomputers in the world.
Here at DeepCloud AI our focus is on building decentralized cloud for decentralized IoT (Internet of Things) and dApps, which lend themselves to running in an decentralized environment. IoT applications require edge computing close to the source, something vastly different from today’s cloud computing paradigm which has a bigger focus on scaling the service or offering consistency. Like the case of specific license plate detection for tracking a criminal using the video camera on traffic lights as they are running in a high-speed chase. Such use cases are better handled by doing computation and detection at the edge and only sending the compressed insights back to the central servers, instead of flooding the networks with live video streams which can choke the entire neighborhood network. We enable resource providers close to the edge, such as retail shopping malls or apartment complex residents, to share excess capacity of their computer resources on the decentralized cloud, close to the city traffic lights, making it possible to do these local computations close to the source and enable such use cases. With our AI matching engine, we match the right resources to the right applications based on real-time analytics of the data across the network.
Other big problem for IoT and dApps is micropayment solution that is cost effective and viable to manage the machine-to-machine and peer-to-peer automated transactions on the shared blockchain infrastructure. Using our Sidechains and Decentralized Acyclic Graph layer on top of the existing blockchain we provide a mechanism for faster transactions with zero fees for transactions.
The lack of good marketplaces for reusable components and tools is also directly addressed by our Application Marketplace, where we provide a pre-validated set of reusable applications and components that dApp developers can use to quickly build their applications.
With DeepCloud AI, we are unlocking the potential of the decentralized application developers so that they can focus on building the core logic for their applications and leave the management of the infrastructure and right tools and components to us.
We believe in truly democratizing the internet and levelling the playing field for cloud computing by opening up the cloud resource market to common people like you and I and provide a spot marketplace where anyone can rent the spare compute resources and benefit from it.
As stated earlier, we will embrace competitors in our space, as it with competition that all of us will get better and be able to take on the larger players like Amazon AWS, Google Cloud, Microsoft Azure. We believe that our value proposition and offering is very much needed and will offer a healthy, improved and powerful alternative for both the overall markets as well as sub-segments of the market.
Come join our community on Telegram and let us know any questions you might have — we are always happy to engage with the community! There will be many more news to come and progress will be constantly shared with everybody as we get closer to releasing our product. See you all at t.me/deepcloud_ai
Geeta Chauhan
CTO, DeepCloud AI
https://www.linkedin.com/in/geetachauhan/
Join DeepCloud AI’s every growing Community:
Official Website: https://deepcloudai.com
Telegram: t.me/DeepCloud_AI
Telegram ANN: t.me/DeepCloudAI_ANN
Twitter: twitter.com/DeepCloud_AI
Linkedin: https://www.linkedin.com/company/deepcloud_ai/
Email: enquiries@deepcloudai.com
