Introduction and the 'killer app'
By 2020 at least 25 billion devices will be connected, according to Gartner. Welcome to the Internet of Things, the next major wave of computing that will supercharge businesses by taking small data from remote devices and collectively making it big data on the cloud.
"To accommodate this growing demand for connectivity, the existing cloud infrastructure will need to grow exponentially," says Jonathan Wilkins, marketing director of industrial automation components supplier European Automation.
"There is a good chance these connected devices will not have internal storage of their own, so will need to use the cloud instead," says Wilkins. "I think it's fair to say that in the next five to ten years we will see 95% of companies use some kind of cloud technology – public or private."
Is the IoT the 'killer app' for the cloud?
"Big data cloud services are the behind-the-scenes magic of the internet of things," says Michael Connaughton, Director Big Data, EMEA at Oracle, who thinks that expanding cloud services will not only catch sensor data, but also feed it into big data analytics and algorithms to make use of it. "The key to competitive advantage in IoT is not just capturing data from the real world, but getting insight before your rivals do."
The IoT is what the cloud was built for – it was designed to connect and collect data and devices from disparate platforms. At the IoT grows, that means ever more disparate operating systems need to communicate.
"Although we have seen a huge growth in cloud-based services over the past few years, as users have transitioned from the use of mostly Windows-based systems to other platforms, such as mobile and Mac, the explosion of IoT devices will increase this growth," says Adam Tyler, Chief Innovation Officer at identity protection and fraud detection technology company CSID.
Could data from IoT sensors overwhelm the cloud?
The IoT will swell the cloud, there's no doubting that – Cisco predicts that annual global data centre IP traffic will reach 10.4 zettabytes by the end of 2019, up from 3.4 zettabytes in 2014, and that's partly because of the growth in IoT devices. Will the network cope?
"The growth of software-as-a-service (SaaS) and the adoption of IPv6 will play an important part in ensuring tomorrow's cloud infrastructure is ready to accommodate the huge increase in data storage, transfer and analytics," says Wilkins. "The IoT can only happen through cloud, but today's cloud infrastructure still has a long way to go before it can accommodate for the huge amounts of data that connected devices will generate."
That's true, but it's incredibly unlikely that data from IoT sensors will ever overwhelm the cloud. "We may see some early issues with scaling, but it is our belief that IoT data will be able to be handled by existing cloud service implementations," says Tyler.
Velocity, not volume
Despite the sheer amount of data continuing to rise, it's important to remember that each sensor will be delivering tiny amounts of data to the cloud.
"It won't only be the sheer volume of the data, but the velocity and variety of the data that the cloud will have to manage and deal with," says Matt Davies, Head of EMEA Marketing at Splunk. "The data coming off a car, plane, train, industrial machinery, production line or building will be used to make critical decisions in real-time, and will need to be combined with other data, some of which will be historical."
Security and challenge of automation
Competing communications standards
Cloud providers need to ensure that they are up to these demands, but some think that it's how each sensor's data reaches the cloud that is more of a problem than cloud capacity.
"The fact that we currently have several radio standards available for connecting servers to the cloud is good news," says Wilkins, "but it is sometimes difficult to choose the best approach for a device or business."
That choice could be cubical, especially if there's to be an industry-wide standard for how fast data is changed between sensors, devices and the cloud. "You don't want to know your car is about to break down and needs predictive maintenance two minutes – or even two seconds – too late," says Davies, who thinks that the same kinds of demands now placed on very low latency, on-premise trading infrastructures in banks will soon be applied to all cloud environments.
How secure will IoT cloud services be?
The IoT may be happening now, but no-one gets to change all of their IT at once. "Insecure platforms, tiny real-time operating systems, and legacy protocols," is what Steve Dunbar, Internet of Things Commercial Director at Microsoft UK thinks are the main security concerns for the IoT.
"IoT cloud services must integrate into environments with devices designed and deployed a decade or more apart," he says, adding that there is an urgent need for novel network architectures that seamlessly integrate them. "Digital security will be increasingly interwoven with physical safety of life and equipment."
Will we see an 'IoT Cloud'?
Given the real-time, low-latency needs of some IoT devices – such as connected cars – there is an argument for a special kind of cloud. "We may well see an IoT cloud, but there will be different tiers within that," says Davies. Any such platform needs to offer analytics and the ability to build new kinds of apps on top of the IoT data.
"The IoT cloud will also need to be able to deliver security to make sure these connected assets are protected, monitored and safe from attacks," says Davies. "An IoT platform will need to learn, predict what is going to happen next, and present this in the form of analytics, alerts and suggested next steps."
Machine learning will also be a core feature of IoT platforms as they evolve, with the data feeding the learning as well as the analytics.
The challenge of automation
Next up for the IoT is automation. "Now that the IoT is expanding beyond the hype into real business applications, the biggest challenge for providers will be the jump from monitoring and analytics to automation," says Nigel Stevens, UK MD of IO.
Describing it as a 'coping mechanism', Stevens thinks that automation is largely about competently handling the sheer volume of data the IoT will generate.
"The amount of data necessary to facilitate M2M communication, automated systems and complex analytics will require modern data networks to spread their storage and computing loads across numerous, global locations," says Stevens. "Efficient, automated platforms are the only way to make the IoT manageable and useful on an industrial scale."
- Also check out: Cloud on clouds: How weather data is a test-bed for new tech
http://ift.tt/23nrksy
No comments:
Post a Comment