IoT's great divorce: Decoupling data from device
The huge variety of systems currently used by IoT devices is holding back the potential impact of big data analysis on companies' business practices. Companies are struggling to gather data from connected devices and transfer it to the cloud or data centres for analysis. However, software designers are seeking to create universal tools, which enable companies to use one program to extract data from a variety of devices. One such firm is VMware, whose product Liota (Little IoT Agent), is a vendor-neutral open source software development kit (SDK) for building secure IoT gateway data and control orchestration applications. This will allows firms to sidestep one of IoT biggest problems: the gateway component.
Why is the gateway such a problem for IoT development?Gateways are an
integral part of IoT infrastructure. We expect the gateway or edge
system to become a very strategic part of the IoT infrastructure.
This is because it not only acts as a bridge, but it also decouples
the physical IoT devices from the risks and insecurities of the
Internet. They also act as the place where connected devices and data
storage components in data centres are aggregated. This bridge allows
data to move freely, correctly and securely from the device to the
cloud or data centre. We are seeing the need for real-time analytics
of the data coming from the connected devices being handled at the
edge or near the edge, while historical analytics are handled in the
cloud or at on-site data centres.
In addition, as IoT
deployment becomes more widespread, organisations want to deploy IoT
apps across multiple gateway vendors. However, there is a wide
variety of gateway development models, making it is difficult for
companies to manage and leverage multiple gateway vendors within
their IoT infrastructures.
Why is Liota important?To understand why,
you first must acknowledge the myriad of gateways, or edge systems.
Think of these as the bridges between connected devices such as
sensors and meters, and the cloud, or a company's data centre. The
problem is that a typical enterprise will have many different types
of connected devices, and they connect to different types of
gateways. If you are a developer writing an IoT application today,
you need to create multiple versions of your application to work on
the different combinations of devices and gateways, making it nearly
impossible to orchestrate across all of them.
How will streamlining the system spur development of IoT and innovation?Developers are often so
bogged down in the day-to-day, backend development of apps that they
rarely get time to focus on innovation. With how quick the market
moves, however, innovation is what will set businesses apart,
allowing them to keep or gain competitive advantage while continuing
to drive the business forward.
Are there unrealistic expectations of the IoT?The IoT is a
different beast. It is not just a question of deploying a typical
application for the company, because for each type of IoT use case,
there are additional physical and infrastructural aspects to be
considered. Each use case requires that you determine the right
sensors, the right edge gateway, identifying the data you want to
use, and what questions it will answer. In addition, because of
privacy, security, latency and bandwidth issues, in many cases,
enterprise IoT solutions will likely not be public-cloud based. The
real-time analytics will need to be handled as close to the edge
gateway as possible.
How is the current situation holding back large scale projects?The current
situation is fragmented because each enterprise is struggling to get
the data out of each device and writing their own one-off extraction
applications. Unfortunately, when they go from "proof-of-concept"
to scale, it becomes very difficult to manage and maintain. This is
where we think we can help to provide a more standard approach to
setting up and deploying IoT applications. Each use case is unique,
but the infrastructure set-up does not have to be unique and can be
standardised across use cases.