Image source: boingboing.net
The dominance of AWS has drawn criticism from some quarters. Following an incident in 2017 when AWS cloud servers went down, many popular services were affected including Slack, Github, and Medium. This led to much speculation about the risks of having so much control over the internet in the hands of one company.
As Wired pointed out
: “The 'winner takes all' dynamic of the tech industry concentrates more and more power into fewer and fewer companies. That consolidation has implications for competition but also affects the resilience of the internet itself.”Not just a risky business
Centralized control over computing power creates more complex issues than risk alone. It is also a necessarily inefficient setup. A cloud computing vendor such as AWS must ensure that it has sufficient spare capacity to scale and meet future demand, as well as assuring performance during peak load times.
This unused capacity means that during any moment in time, there is a large amount of computing power that is idle. For any computing vendor, this is unprofitable in the same way that having unsold physical inventory sitting around is costly for a company selling goods.
The ‘Big Four’. Image source: LinkedIn
As demand for computing power is increasing, there is currently no other option for vendors of computing power, if they are to ensure they can service new customers. One of the most significant factors driving annual growth over 20%
is the developments in artificial intelligence (AI).
AI is being driven by the availability of big data, which feeds into machine learning algorithms. Crunching such enormous volumes of data requires a lot of computing power. Obstruction to innovation
Access to the necessary computing power is where smaller AI developers struggle with bringing their products to market. Early stage startups need to do extensive QA testing during the R&D process, often rendering projects prohibitively capital intensive to get off the ground. The options for an AI developer are either to purchase their own serversor to buy a cloud computing contract.
Cloud computing is usually the lower cost option of the two. However, the market domination by a small number of big players means that the cost of computing is still prohibitively high.
For this reason, many of the top AI companies
are the same companies as those providing computing services—Amazon with its Alexa robot, Google’s Deepmind, and IBM Watson. Other big names in AI development are also big tech companies with big budgets to match, such as Facebook, Twitter or Apple.
Smaller companies need significant continued investment to compete, which can be difficult to secure if returns will only happen far in the future. Unblocking progress with blockchain
Many of the problems and barriers to computing services arise as a result of a heavily centralized marketplace. Therefore, the decentralized power of blockchain could prove to be a winning solution for the future of cloud computing services. Tatau
is one company that is seeking to commoditize compute power by creating a blockchain-based computing platform.
Tatau believes it can become “the Uber of compute.” By leveraging its blockchain-based network of computing resources, AI developers—or anyone else needing computing resources at scale—can tap into this platform at a lower cost and on an as-needed basis. The company will be releasing its beta-version early next year.
Using blockchain in this way overcomes many of the shortfalls of a centralized marketplace. For smaller AI developers, it provides unprecedented access to more cost-effective compute services needed to bring their projects to market, without being tied into lengthy contracts.
It also offers similar flexibility to the big cloud compute vendors like AWS. Where a vendor finds it is operating under capacity, it could put this to use through a blockchain platform by selling it on an as-needed basis. If they later need the server capacity for one of their own customers, they can simply withdraw it from the platform when they need it.
Finally, decentralized computing power is less prone to downtime than centralized servers. Because the compute comes from a distributed network, other nodes can cover the work of any node that may be down for a period of time. Additionally, decentralized networks are less prone to malicious attacks. Powering the future
Emerging technologies such as AI and blockchain are creating the next technological revolution. It is vital for the ongoing development of these technologies that innovation is fostered, and not thwarted.
By breaking down oligopolies on the necessary infrastructure such as computing power, smaller developers can have access to the resources they need. At the same time, if the big vendors can also profit, it’s inarguably a win-win scenario.