In-depth Software Development Information

Twitter: @ericjbruno

A Cloud Approach for the Internet of Things

Eric J. Bruno
October 2, 2015

Extend Your IoT Data Reach with the Cloud
The Internet of Things (IoT) is part of a promising future where smart computing devices will communicate with one another, with people, and with the enterprise applications we use. But IoT value is in the data, reaching from the enterprise out to realize the tremendous potential at our fingertips. This data-centric view of IoT requires the integration of new data services, the injection of intelligence into our devices, and the interconnections with enterprise systems.

Soon, we’ll see growing numbers of devices and systems connected in ways we haven’t yet imagined, requiring us to handle a wider variety of data types with even greater volume and velocity. To use IoT to discover what’s truly important and relevant to your business and your users, you need to have the right IT capabilities. The technical requirements and data volumes introduced are best met by the cloud, which surpasses traditional data center implementations in multiple dimensions.

Devices and sensors are a means to acquire valuable data, and IoT applications often require a distributed set of services to act on that data. The timely, contextual data, along with the cloud, are so crucial to IoT, they essentially transform and extend enterprise architecture beyond what we know today. Although on-premise enterprise application infrastructure has much in common with IoT infrastructure requirements in terms of technology, it’s the cloud that solves many of the challenges head-on.

IoT Challenges and Barriers to Entry
Up until more recently, one major challenge enterprises faced when trying to deploy an IoT project was configuring and integrating the myriad of devices and the numerous—and sometimes competing—protocols required. While devices and their challenges receive a lot of focus in IoT, there are other challenges that are just as important. These challenges include the need to deploy dedicated servers, middleware, databases and analytics packages for an IoT application, with potentially different communications solutions for each device type used. This accounts for a great deal of infrastructure and cost, potentially multiplied by each IoT solution deployed.

Other challenges include systems integration, security, data aggregation, data management and analytics, and scalability and elasticity.

IoT Business Concerns
As with any technology implementation, the motivation behind IoT applications begins with business drivers and concerns. First, you need to identify the business of IoT and focus on customer value and return on investment (ROI). Part of this process is identifying how to deliver value quickly, to both your business and your customers.

Here's a set of business concerns for any IoT application, which need to be directly addressed in the solution:

  • Security and reliability: all data needs to remain secured, unchanged, and persisted at each stage of an IoT solution. This includes distributed components, communication infrastructure, back-end analytics and database servers, across potentially remote locations and adverse environments.
  • Data acquisition: data is the lifeblood of every IoT application, requiring reliable collection, security, and delivery to back-end processing server.
  • In-flight Processing: you often need the ability to operate on data as it’s collected in real-time, and then store and combine it with historic data to mine for additional value.
  • User and device provisioning: data is collected from embedded devices and sensors in the real world on behalf of people or machines. Therefore, you need to efficiently provision devices as part of your application, associated with your authorized end users. This is a crucial yet overlooked requirement for IoT solutions.
  • Cost of first entry device: the time, effort, and cost associated with deploying your first device or IoT application can be daunting. Reducing the associated costs, effort, and risks can help you experiment and find value in IoT.
  • Ease of deployment: you need the ability to configure, update, and deploy entirely new applications to your system of dispersed infrastructure quickly and easily.
  • Robustness: more can go wrong with a solution that’s less centralized. You need a solution that’s elastic and scales appropriately to avoid failures, and is better suited to deal with unexpected events.
  • Innovation: the underlying motive is to be as innovative as possible without allowing existing technology and processes to get in the way.

IoT and Interconnected Services
The next step is to identify the top priorities, beginning with what’s most important to customers now while preparing for the future. For example, outline a growth strategy for your IoT solution, understand how to respond to critical problems such as communication and power failures, intermittent connectivity, and ultimately how to improve the solution with fixes and added functionality.

It helps to have an agile, iterative process to IoT development, where the most important items are built and delivered early, with user feedback gathered along the way. The reward is in the ability to learn quickly and make minor course corrections sooner, and more often. The challenge is finding a development platform that can support this without overwhelming cost and effort. Let’s examine how the cloud can be the best choice for IoT.

The Advantages of the Cloud for IoT
IoT is a trendy technology topic, and most agree it holds a lot of promise. But there’s a risk that one-off IoT solutions based on existing on-premise IT cost too much to support proof of concept projects, take too long to deploy, and threaten to stifle innovation. According to IDC, within the next five years, over 90% of all IoT data will be hosted on service provider platforms as cloud computing reduces the complexity of supporting IoT "Data Blending".

The cloud is one answer to achieving an agile and iterative development platform, further enabling you to achieve the promise of IoT. For instance, the cloud is always there and always on, reducing development and deployment bottlenecks (think DevOps). It does this by providing the resources and capacity needed, with reliability, operational efficiency, and near unlimited bandwidth.

In general, the cloud can deliver a global, nimble solution base, lower maintenance overhead, best-in-class security, and an overall reduction in risk. It’s also flexible, cost effective, scalable and elastic, and can decrease time to market. A quick comparison between the IoT business concerns listed above and the value of the cloud reveals that they’re a near-perfect match. Let’s take a closer look at the benefits of the cloud and how it matches the specific business needs of IoT.

The IoT Cloud Business Advantage
With visibility into your entire set of services, a cloud provider can properly secure and optimize traffic to meet your global bandwidth needs. The cloud can further improve uptime and performance through adaptive traffic management, protocol and data normalization (for the varieties of IoT device communication and data), deep application visibility, and dedicated focus on quality of service (QoS) agreements.

Here are some direct advantages the cloud offers IoT:

  • A unified security model with identity management and authentication services to comply with regulations such as PCI and HIPAA, often with an SLA or QoS contract to help assure the security and compliance of your data.
  • The elasticity needed to meet spikes in demand. The cloud provides scale to handle periodic large volumes of data without building out your own large, dedicated IT infrastructure needed to support it. This means you can affordably and effectively handle both occasional and frequent spikes in data demands from rapidly scaling devices in your IoT applications.
  • Accessibility for connectivity on a global scale via geographically distributed datacenters with optimized routing and caching implementations for the best bandwidth and request/response times possible.
  • Flexible storage options for the large volumes and varying types of data that will grow to fit your needs.
  • Unified event processing across real-time and historic data sets.
  • A set of deep data analytics, all in one place accessible across solutions in support of a single source of truth.
  • Global application, user, and device management from a central location
  • A standard application hosting and deployment model (i.e. DevOps) with a unified access layer across web and mobile applications.

The Cloud Improves Economics
With a pay-as-you-go model of buying cloud resources—instead of the large capital expenditures associated with building out your own datacenters—you pay only for what you need when you need it. Cloud subscriptions move you to an amortized model for IT expenses. As a result, you pay for an outcome (with an SLA) as opposed to technology resources. Add to this the savings of outsourcing monitoring and support, improved tools and processes for software deployments, and layers of virtualization and elasticity provided.

There are other benefits of not operating your own IT datacenter, such as savings in power, cooling, and telecommunications costs. These non-essential business items are deferred to a cloud provider who specializes in offering this as a service, with related management and expenses spread out over many of its own customers. As a result of these changed economics, IT resources mutate from something that’s rationed, into a set of services that are effectively unlimited yet affordable.

The IoT Cloud Security Advantage
Security is often at the top of the list of most organizations’ IoT challenges. This includes the perceived risks involved in allowing your cloud provider to become the caretaker of your information. Others back this view, illustrating how security continues to be at the top of the list of concerns for IoT, both for organizations building solutions and their users. For some, IoT security is so much of concern that entire studies have been conducted on the topic.

Depending on the capabilities of the provider you chose, an enterprise could actually improve the overall security compared to what it might otherwise be able to resource on its own. For example, the centralization of the cloud makes it easier to integrate device and user identity into an IoT workflow. Existing enterprise identity security can be leveraged for devices to solve security issues around authentication, authorization, and access control is a uniform way. Additionally, some providers offer advanced security threat analytics based on the large volumes of data and traffic they’ve handled for years.

The top five sets of security-related questions you should ask of any cloud service provider:

  • Data transparency: Where does your data reside? Does the provider have a dedicated security group? Are there independent security audits? Is someone accountable for data security?
  • Risk mitigation: Are certifications and third-party validations performed? How is end-user access managed and revoked, if needed? Is a secure, centralized single sign-on solution offered?
  • Proof of capability: Can security assertions be proven? Can security and regulatory compliance be demonstrated?
  • Integration: How will your existing enterprise applications integrate with cloud-based services? Can your public cloud applications securely access data in your private cloud or on-premise solution?
  • Amount of experience: Which vertical markets does your vendor support? Is there a consistent implementation in terms of data controls across markets?

The IoT Cloud Data and Analytics Advantage
Cloud-resident technologies enable the efficient collection, storage, preparation, and analysis of all the data assets of an organization. With the rapid adoption of software-as-a-service (SaaS) applications and the accumulation of IoT data at unprecedented rates, most of it has also moved to the cloud. This is known as data gravity, which implies that as your data gathers in one place, supporting services are often built around it, and vice versa.

With this shift of data gravity, powerful yet easy to use cloud-based analytics has become paramount. The reasons are several:

  • As data and services grow larger, they become more difficult to move. Building out your IoT solutions in the cloud now will ensure you have a scalable solution to meet growing demands moving forward.
  • Data preparation (without coding) needs to allow users to find, profile, explore, and quickly transform the myriad of data points into ways that can be navigated intelligently to discover root causes, uncover trends, and predict future outcomes. The cloud offers a unified repository to make this possible at scale.
  • Business analysts, data scientists, and IT staff need to be able to collaborate on IoT analytics and Big Data projects. It’s essential to provide the data and services centrally, in the cloud, to facilitate this.
  • Integrated with analysis capabilities such as enterprise social graphs and cross-channel information, you can perform pattern recognition across enterprise transactions and tie those to interactions from social media and social networks to determine successful (as well as negative) business patterns.
  • Search-based data discovery tools, together with predictive analytics, operating on your centralized cloud-based data set, if done right can provide you the ability to not only see a problem before it occurs but also to take action to correct it, giving you "actionable analytics".
When hunting for the value of IoT, true discovery comes from exploration. To do this, you need the ability to deploy multiple IoT applications quickly, where some may even overlap. Building an on-premise solution for each application would be cost preventative, but with the cloud it’s as simple as an experiment. That’s the power of the cloud, and how it will unlock the value in IoT. True business success is all about service-level agreements, not about the bits and bolts of server infrastructure. Here are some examples of how IoT solution SLAs may vary from standard datacenter SLAs:

  • IoT applications have varying needs for message transmission between devices and enterprise systems. Some require bounded latency, while others require minimal loss rate For IoT and cloud solutions, SLAs need to emphasize operational health and business value, not just technology-related measurements
  • Measurements need to include the speed in which new services and entire applications can be deployed or delivered
  • End-to-end response may need to be more flexible, and adjusted based on physical world disruptions and environmental factors. You may also require built-in support for intermittent connectivity.
  • For cloud providers with shared infrastructure, the cloud vendor needs to assure you that market changes that affect other clients’ applications won’t impact yours.
  • IoT solutions require assurances that even with data and usage spikes, associated IT costs will remain lower with a cloud solution
  • Mobility and social media drive user demand around the clock, as users connect anytime and from anywhere, driving true 24x7 customer service requirements
IoT and Cloud Benefits Across Lines of Business
For lines of business, such as retail, telematics, healthcare, and so on, IoT promises specific benefits. For instance, IoT can improve retail marketing, where analytics and augmented beaconing data can identify shoppers and send relevant offers to them in context with their location and preferences. In telematics, diagnostics of a vehicle’s embedded systems can be used to predict and prepare for maintenance in a timely manner, before breakdowns or safety hazards occur. In healthcare, patients can be monitored remotely and global treatment outcomes can be compared and then predicted according to patient data. For building automation, enhanced IoT based monitoring can help reveal if you’re spending too much on electricity, predict future power needs based on patterns, leading to more efficient electricity usage.

In each of these cases, the cloud helps provide distributed application provisioning, end-user remote device and data management, security and regulatory adherence, centralized data storage and lookup, and comprehensive analytics. In addition, the cloud is critical in providing global communication channels for the wide rollouts needed in IoT solutions. Arguably, no on-premise IoT solution will offer a cost effective, complete, and robust solution like the cloud.