How Blockchain can Create the World’s Biggest Supercomputer
As our desktop computers, laptops, mobile devices, etc. stand idly by for a huge portion of the day, the need for computing resources is growing at a fast pace. Large IoT ecosystems, machine learning and deep learning algorithms and other sophisticated solutions being deployed in every domain and industry are raising the demand for stronger cloud servers and more bandwidth to address the minute needs of enterprises and businesses.
So how can we make a more economic and efficient use of all the computing power that’s going to waste? Blockchain, the distributed ledger that’s gaining traction across various domains, might have the answer to the dilemma by providing a platform that enables participants to lend and borrow computing resources — and make money in the process.
The rising challenges of computing
“There is a growing demand for computing power from industries and scientific communities to run large applications and process huge volumes of data,” says Gilles Fedak, co-founder of iEx.ec, a distributed cloud computing platform.
Fedak names several domains, such as product simulation, deep learning and 3D rendering, where demand for expensive computing resources and High-Performance Computing (HPC) is rising.
“The biggest challenge for supercomputing is the demand to compress time,” says Jerry Cuomo, vice president of Blockchain for Business at IBM. “Business processes must now be completed at a significantly faster pace than before. The result is that the demand for computing power is increasing exponentially.”
David Sønstebø, founder of IOTA, a distributed ledger for IoT, also underlines the need to achieve real-time computation and overcome the lag caused by current cloud-based models. “The biggest problem for computation overall is that the devices generating data are not located close-by to the data centers that perform the analytics,” he says.
How distributed computing solves the problem
Compute resource sharing platforms such as SETI@home have existed for years. But they still depend on central brokers to distribute and manage tasks, which can make things complicated.
One of the fields where centralized and cloud-based computing falls short is the Internet of Things, Sønstebø says. “As IoT grows, the need for distributed computing becomes an absolute necessity,” he says. Latency in round-trips, network congestion, signal collisions and geographical distances are some of the challenges faced when processing data produced at edge devices in the cloud. “Devices need to be able to trade computational resources with each other in real time so that the computational load can be distributed,” he says.
Some of the emerging lines of software will not be supported by centralized architectures at all, iEx.ec’s Fedak says, such as decentralized applications (DApps), which, among others, will power fog computing, distributed AI and parallel stream processing. “This class of application is extremely challenging because they’re both data and compute-intensive, and they don’t cope well with centralized infrastructure,” Fedak says.
Incentivizing resource sharing is also a problem with centralized models.
“If you look at the last 10-20 years’ of progress in virtualization, it’s obvious that setting up any kind of environment in a data center or on an individual computer has become much easier,” says Julian Zawistowski , co-founder and CEO of distributed computing platform Golem. “But when it comes to actually renting the hardware, it still tends to be painful: comparing the offerings of different providers is complicated, and it takes quite a bit of time and expertise to figure out the best solution for a given task.”
“The issue with getting payment involved is that you need to check whether the participants are actually performing the work and also integrate payment so that the provider of the compute capacity knows that running the computations is going to be worth its time,” says Preston Byrne, COO at Monax. This is easy when you’re dealing with trusted entities such as the Amazon Web Services HPC platform, but not so when you’re dealing with nodes that vary in hardware and power.
…
Read the full article written by Ben Dickson on TechChrunch.com
Read full article