April 20, 2017

Why Enterprises Will Need an IoT Operations Center

By Special Guest
Manu Tayal & Arsalaan Kashif, Manu Tayal, general manager; Arsalaan Kashif, associate director of product engineering services, Happiest Minds Technologies

IoT has been a buzzword for a few years. But it’s only now that organizations are moving from the proof of value to the productization stage of their IoT journeys. As things pick up steam, it is only natural that a number of organizations will soon initiate enterprise-wide rollouts. That will require an increased focus on operationalizing IoT.

So far, organizations have focused on certain elements of IoT that have caught their fancy while glossing over certain other crucial aspects. However, they will soon have to connect the dots for an elevated IoT experience that can be sustained over a long period of time.

This is where a dedicated IoT operations center becomes critical in ensuring a smooth and seamless enterprise-wide IoT rollout. This kind of center can ensure improved connectivity, increased control over devices, and effective cloud and infrastructure management while allowing for high levels of security and customer support.

This is how an IoT operations center can stitch together the various elements that go into enabling a smart, secure, and connected IoT experience.

Connectivity Management
It is widely estimated that in the next 5 years, there will be 21 billion things that will be connected across the world. It’s safe to say that the need for effective connectivity management cannot be overemphasized. In a world where smart cars are communicating with parking sensors, a disruption in connectivity can prove to be disastrous.

However, in the event of compromised connectivity an IoT operations center would be able to use the right tools to ascertain the reasons for failure and take the appropriate steps to remedy the issue.

Monitoring connectivity from sensor to gateway and gateway to cloud Sensors to Gateway: The connectivity could be wired or wireless and when the data from a sensor stops, there is a need to identify the faulty sensor and ascertain the cause.

The cause could possibly be due to one of the following reasons:

The sensor has gone bad.
Fix: Replacement of sensor

There is no change in value and hence no data is sent.
Fix: Not required

Interference has been introduced.
Fix: Ask the field crew to change the environment.

The interface on the gateway collecting the data has gone bad.
Fix: Replacement/fix the gateway

Gateway to Cloud: The connectivity could be cellular, Wi-Fi, or satellite.
The possible reasons that need to be looked into:

Has the device gone into an area where the signal was supposed to be weak?
Fix: No fix required. This seems to be a movable device. If it supports a store and forward feature, there wouldn’t be any loss of data.

Has there been an issue because of non-payment and hence the connectivity has been disabled by the service provider?
Fix: Notification to make the payments

The gateway is malfunctioning.
Fix: Troubleshoot the gateway

Activation and deactivation of connectivity
Based on the payments made by the customer and the type of subscription availed, the operations center should be able to disable connectivity (SIM in case of cellular connectivity). Also there may be instances where the SIM card is being misused and the operations center wants to prevent misuse by disabling connectivity.

Metering the data usage
It is crucial to meter the usage of bandwidth for various purposes such as cost distribution across different managed devices. In case this is a service offered by an OEM or a service provider, this data will be used to generate a bill for the user. The different types of usage that need to be monitored are:
• the number of devices being managed;

• the amount of data being exchanged by the gateway to the cloud platform; and

• the amount of subscribed value that has been utilized by the user. This is measured via a sensor or a combination of sensors on the field.

With the plethora of devices that will now be connected to each other, it’s going to be an onerous task to ensure that all of them are working cohesively and in harmony with each other. Even a minor malfunction could throw a spanner in the works and send the entire operation spiraling out of control.

As a lot of these devices would be deployed far and wide, the operations center would need the right tools to manage these devices in the field. These devices would go through multiple stages in their lifecycle and would need to be managed through each stage.

Inventory management, commissioning, and replacement of the devices or sensors
Devices need to be procured or manufactured and tested for functionality before they are commissioned into the field. More so if they have to be fitted with a SIM card, as the association between a SIM card, device ID, and the end user has to be established before the device is dispatched for installations.

Maintaining this inventory along with some spares for faulty devices is the key to smooth deployment. The device also has to be loaded with some certificate or key, which will be used for authentication once the device goes live in the field. The commissioning of the device ensures that the devices are installed at the user premises and securely connected to the backend cloud platform. The credentials are checked against the already provisioned entries of the device in the backend.

All of this requires supply chain management and hands and feet in the regions where the rollout is happening. If the rollout is nationwide, for example, it is important to create systems for device replacement. There could be regional locations where inventory for device replacement is maintained. This helps in quick replacement and avoids shipping delays.

Configuring and controlling the gateway
Most of the time, the devices will be deployed in the field with a default configuration. This may need to be changed based on the specifics of the customer, location, and use case. The operations center would need to send that configuration down to the device remotely.

Monitoring the health of the devices and calibration of the sensors
One of the KRAs of the operations center is to ensure minimum downtime of the devices. Also it is important to keep an eye on a possible breach of security. Typically, devices send the heartbeat signal to let the system know about their status. If not, specific heuristics may need to be developed to detect any anomaly in the device behavior to detect device problems.

The other aspect of importance is calibration of sensors. If the business decisions are going to be taken based on the sensor data, it better be sending the right data. Each sensor has its own life, and it’s important for the operations center to have the tools to predict the life of a sensor with reasonable accuracy.

Upgrade the software on the devices at module, application, and system levels
Bugs exist. They need to be patched. Features evolve. They need to be upgraded. All of this can happen at various levels – module, application, or system.

The operations center needs to be responsible for patching and upgrading the devices. The key is the time of upgrade and the level at which it happens. Upgrading the whole system when there is a small application fix can be cumbersome. Obviously it depends on the levels of upgrades supported by the devices.

Cloud Infrastructure Management
The quantum of connected devices we are likely to see in the future will place tremendous demands on traditional architectures, causing them to eventually crumble under pressure.

In addition, challenges with scale, reliability, and a faster turnaround to business demands will prompt a move to a cloud-based environment. This is where the IoT platform and the business applications will be hosted. Although the devices may not be connected all the time due to connectivity charges, the cloud infrastructure would typically be expected to be up and running 99.X percent of the time.

The dedicated IoT operations center would play a key role in addressing the following aspects.

Monitoring for scale and performance
The speed at which the scale and performance of the cloud infrastructure can change can be much higher than the traditional IT infrastructure. It is very important that as the infrastructure is set up with minimum available resources, it is monitored round the clock.

Two factors that need to be constantly looked at are data ingestion volume and storage capacity. It is important to have the right data aggregation and archiving strategy in place to run the business at optimal cost.

Monitoring for uptime
This is typically the SLA signed up with the customer. It is important to ensure adequate amounts of highly available infrastructure along with the right DR site. This is where most decisions pertaining to business criticality and cost are taken.

Dynamic support for new protocols and new devices
As new devices are added to the infrastructure, there would be a need to support new protocols. There would be device-facing components on the cloud that would need to be added for proper communication with end devices. Also the data would need to be mapped properly for the right storage and analytics. The dedicated operations center will be equipped with this in the long run and have an understanding of the platform and related devices at all times.

Security has been a major concern universally and is likely to climb up the charts with the number of connected devices on the rise. In an environment that has a wide range of devices connected to each other, vulnerabilities can creep in rather easily and compromise the entire system. Hence, security must be the foundation of any IoT rollout. But currently there is no standard way of securing an IoT device.

Monitoring device security
The internet is still not secure. Devices have only added fuel to fire. The biggest challenge is that limited processing power restricts encryption and other robust security measures. It is very important that the operations center ensures that the devices authenticate at a minimum before they get onto the network.

Many of the mission-critical systems are looking at only monitoring the data rather than controlling the devices as a way to limit the damage that might happen. It is important for the IoT operations center to understand the limitations of these devices from a security standpoint and have systems in place to detect a problem early. The operations center needs to be constantly looking out for security patches for the devices from the device vendor and the embedded software vendor and apply them at the earliest opportunity.

Monitoring cloud security
This aspect of security has matured a bit over the years. But it continues to be something to watch out for, especially on the side of device connectivity. However, it is important to do a vulnerability assessment of the setup.

Monitoring application security
Application security and specifically API security, also need to be taken into account. As IoT evolves into the internet of businesses, the API economy is likely to become a major force. It is important that the operations center has the right toolset to monitor API access and the security around that.

Customer Service
Even though IoT is powered by things, it is ultimately all about ensuring the best possible experience to a customer and elevating that customer’s quality of work and life. Even though all of the elements outlined above are key to an operations center, it’s real value lies in ensuring the highest levels of customer service. Toward this end, the three main things the IoT operations center should focus on are:
Insights into operational performance to the customer
Insights into how efficiently the whole IoT infrastructure is functioning across devices, cloud, connectivity, and security are critical. It is important to get back to the customers on insights based on the usage of the device to suggest optimizations in configurations.

These optimizations need not be limited to the use cases which have been IoTized, but also the use cases which if IoTized will enhance the operational performance of the system.

Insights into cost performance
Once deployed, the biggest recurring cost for the end customer is the connectivity charges and cloud infrastructure charges. It is important for the operations center to keep tabs on both. Based on the usage of data, if there is data aggregation that can be done at the gateway, or if the frequency of data transmitted can be reduced, it will amount to significant savings in connectivity costs.

On the cloud infrastructure side, it is important to set up systems to scale them up and down as needed. It is important to identify the minimum viable infrastructure for the current scale of deployment and scale it up when the need arises.

Insights into product performance
This is the aspect that’s of utmost importance to OEMs. As the data gets collected at the operations center, the operations center can work with the enterprise and the OEM to set up a system to provide critical insights to the OEM. This will help the operations center to service the end customers better, while also effectively monetizing the data that has been collected.

Overall it is inevitable that the concept of a dedicated IoT operations center becomes a reality as IoT rollouts become increasingly common. Yet most organizations that are embracing IoT aren’t fully geared up to embrace all aspects of an operations center, which can very easily lead to sub-par results.

Manu Tayal is general manager and Arsalaan Kashif is associate director of product engineering services at India-based digital transformation company Happiest Minds Technologies (www.happiestminds.com).

Edited by Ken Briodagh

Back to Homepage
Comments powered by Disqus