Jon StojanCloud computing, or the on-demand availability of computer system resources such as storage and computing power, has been a huge boon for individuals and organizations all over the world. Centralizing computing and storage resources, similar to public infrastructure, has significantly reduced the cost of maintaining IT infrastructure for businesses.
Previously, each company had to invest in expensive on-site servers and have a large IT team to maintain everything. Today, with cloud computing, businesses no longer need to purchase as much equipment, and IT teams can be much leaner, as most of the responsibility for maintaining infrastructure is handed over to the cloud service provider, which include tech giants such as Amazon, Microsoft, Google, and Alibaba.
However, with more and more businesses “hopping on the cloud”, issues regarding reliability have emerged. Service interruptions of major cloud providers have made headlines for their wide-reaching effects. On February 28, 2017, Amazon Web Services suffered an outage for several hours, affecting a huge number of businesses’ operations. More recently, the July 2024 Microsoft outage caused interruptions in many sectors, including air and rail travel, across many countries worldwide, and the services for the London Stock Exchange.
In light of these issues, Mr. Steven He, founder of cloud computing 2.0 company Ubiq, Inc., says that, for cloud computing to advance and meet the increasingly exacting demands of clients, then there should be a paradigm change with regard to reliability and the architecture of cloud services. The increasing centralization of cloud computing has introduced a single point of failure that could bring entire operations down and greatly jeopardize the global economy.
Ubiq – Logo
Steven explains that a cloud service’s reliability or uptime is measured in “nines”. For example, 99% uptime is “two nines” and 99.9% and 99.99% are “three nines” and “four nines”, respectively. An uptime of four nines means an average of around 52 minutes of downtime each year, which is still too high for mission-critical industries such as banking, aviation, or utilities. These industries often require a reliability of five nines, or an average of less than six minutes downtime per year.
As businesses become increasingly reliant on cloud technology, Steven, who has more than 30 years of IT experience, including in enterprise content management systems, saw the need to create a solution that retained the convenience and benefits of cloud computing while solving the single point of failure issue, increasing its reliability.
His company’s flagship project, UBIQ Network, is a cloud computing platform that features a patented decentralized architecture that can provide a level of reliability that meets its client’s needs, even up to five nines or more. UBIQ Network now has more than 40 nodes worldwide, primarily in the US, China, Canada, Germany and Hong Kong, with multiple software applications now running on the platform. These nodes act as failsafes and provide redundancies in case the node closest to a user goes down.
According to Steven, it can be very expensive to increase the reliability of conventional cloud computing, requiring an exponential increase in investment. On the other hand, due to UBIQ Network’s decentralized nature, offering a higher level of reliability involves adding nodes and it needs just a linear increase in cost, resulting in more savings for clients.
“The recent Microsoft outage’s huge global impact is being compared by some industry members to the damage that would’ve happened if the Y2K issue was not solved prior to the turn of the millennium,” Steven says. “I believe that this is the right time to reexamine cloud computing and its reliability issues. If we don’t fundamentally change the architecture and continue concentrating more computing power and storage resources at a singular point, then an even more catastrophic failure is bound to happen sooner or later.”
Aside from improving reliability, a decentralized architecture like UBIQ Network also solves another problem with cloud computing, which is the lack of data mobility. Because cloud computing concentrates its resources in data centers, inefficiencies in transmission result in poorer performance the farther a user is from a data center. Steven says that this can have a huge negative impact on technologies such as full self-driving cars, which are already being developed. A fully self-driving car needs seamless communication with the cloud, and a delay of even a few seconds can pose a huge risk to the vehicle’s occupants, especially when moving at high speeds.
“The laws of physics dictate that communication quality and speed is inversely proportional to distance,” Steven says. “Which is why a decentralized system is important if technologies such as fully self-driving cars are to prosper.”
Having seen the rapid development of computing technology across his over three decades in the IT industry, Steven foresees a huge shift in cloud computing over the next few years, as more reliability issues come to light.
“In the 1980s, supercomputers dominated the market, and companies sought to buy the largest, most powerful computer they could afford,” Steven says. “However, the miniaturization of computers and emergence of PCs and the internet caused supercomputers to mostly fall out of fashion. Similarly, 10 years from now, miniaturized and geographically distributed data centers will take away the market dominance from today’s large-scale, centralized data centers.”