Architecting the Industrial Internet

Shyam Nath, Robert Stackowiak, Carla Romano

• Publisher: Packt Publishing
• Year: 2017
• ISBN:
  9781787283749
• Edition: 1
• Topic: Ciencia de la computación
• Subtopic: Computación en la nube

Learn the ins and outs of the Industrial Internet of Things through subjects ranging from its history and evolution, right up to what the future holds.About This Book• Define solutions that can connect existing systems and newer cloud-based solutions to thousands of thousands of edge devices and industrial machines• Identify, define, and justify Industrial Internet of Things (IIoT) projects, and design an application that can connect to and control thousands of machines• Leverage the power and features of a platform to monitor, perform analytics, and maintain the Industrial InternetWho This Book Is ForArchitects who are interested in learning how to define solutions for the Industrial Internet will benefit immensely from this book. Relevant architect roles include enterprise architects, business architects, information architects, cloud solution architects, software architects, and others. The content is also relevant for technically inclined line of business leaders investing in these solutions.What You Will Learn• Learn the history of the Industrial Internet and why an architectural approach is needed• Define solutions that can connect to and control thousands of edge devices and machines• Understand the significance of working with line of business leadership and key metrics to be gathered• Connect business requirements to the functional architecture• Gain the right expectation as to the capabilities of Industrial Internet applications and how to assess them• Understand what data and analytics components should be included in your architecture solution• Understand deployment trade-offs, management and security considerations, and the impact of emerging technologiesIn DetailThe Industrial Internet or the IIoT has gained a lot of traction. Many leading companies are driving this revolution by connecting smart edge devices to cloud-based analysis platforms and solving their business challenges in new ways. To ensure a smooth integration of such machines and devices, sound architecture strategies based on accepted principles, best practices, and lessons learned must be applied.This book begins by providing a bird's eye view of what the IIoT is and how the industrial revolution has evolved into embracing this technology. It then describes architectural approaches for success, gathering business requirements, and mapping requirements into functional solutions. In a later chapter, many other potential use cases are introduced including those in manufacturing and specific examples in predictive maintenance, asset tracking and handling, and environmental impact and abatement. The book concludes by exploring evolving technologies that will impact IIoT architecture in the future and discusses possible societal implications of the Industrial Internet and perceptions regarding these projects.By the end of this book, you will be better equipped to embrace the benefits of the burgeoning IIoT.Style and approachThis book takes a comprehensive approach to the Industrial Internet, thoroughly acquainting the reader with the concepts and philosophy of the IIoT. It provides a basis for defining an IIoT solution in a thoughtful manner and creating what will be viewed as a successful project.

Defining a Deployment Architecture

You learned about the information domain components that govern the Industrial Internet architecture. We reviewed the data and analytics requirements of standalone manufacturing facilities and the emerging Lambda architecture in Chapter 6, Defining the Data and Analytics Architecture. We went over the speed and batch layers based on the business needs and use cases. We introduced Hadoop and the concept of the data lake for decision support.

In this chapter, we will look at the different considerations and trade-offs that are a result of the different deployment architecture and strategies. You will learn to fully appreciate that architecture is the art of compromise (https://www.world-architects.com/en/architecture-news/reviews/architecture-of-compromise). The current organization strategy and culture will play an inadvertent part in the deployment architecture for Industrial Internet applications. This chapter will focus on the design of this deployment strategy where the line of business and the CIO organizations may have a healthy friction. A savvy architect needs to understand how to facilitate such debate and discussions between the stakeholders.

Here we will deep-dive into the technology choices and the deployment architecture for Industrial Internet solutions. This chapter will show you the right place to deploy such applications all the way from the edge to the cloud and enterprise systems.

The chapter will covers the following topics:

• Past and current deployment architectures
• On-premises or cloud deployment
• Design for edge tier
• Networking considerations
• Device management
• Management and support infrastructure
• Consumption models

Current state of deployment architectures for IT systems

The architecture of IT systems has gone through several revolutions since early mainframe systems. Mainframe systems handled all data-management, computing, or business logic, and were housed in a very large cabinet. Early mainframes performed data processing in batches and could only be accessed by dumb terminal emulators with no intelligence, and dedicated printers. They had limited support for graphical user interfaces (GUIs) and couldn’t access multiple databases from dispersed locations.

With the introduction of the personal computer (PC), small-to-medium-sized businesses could afford to shift their business functions and data to PCs, where the processing could be controlled by the user. The Early PCs were closed systems limited to a single PC and had limited memory and data storage capabilities, and a PC failure could result in a disastrous loss of data and business processing. Data and computing logic could only be shared via peer-to-peer networking or by making a physical copy of the data to transfer via external storage means, such as floppy disks.

Figure 7.1: Client-server architecture was enabled with the advent of networked PCs and modular programming approaches. With client servers, the database or file server resides in a centralized server, and the application logic is distributed to client workstations or desktops, while the data can be queried and presented using GUI tools such as PowerBuilder and Forms. With the business logic and presentation resources aburden on the client, it was often referred to as the fat client architecture in which each user of the system would interact using a client with the presentation tools and application logic installed locally. This could present challenges for upgrading software, as the upgrade needed to be coordinated among all the clients and server(s) in the system. Database engines added the ability to offset some of the computing capability with stored procedures, and functions were executed by the database engine, relieving some of the fat-client load.

The defining component of client-server architecture was the network. When the client application issues a query, the request is transmitted from the client to the server by the network. The server executes the request and returns the results or data back to the client via the network. Client-server architecture enabled distributed computing, and distributed databases led to more heterogeneity within a system and increased network and connectivity requirements. It also introduced challenges in managing transactions, especially in multi-user systems, and challenges in ownership (that is, user versus central) for hardware, network, support, data, applications, and configuration management:

Figure 7.1: Client-server architecture

The internet provides the ability for systems to communicate across Internet-enabled servers and clients (Figure 7.2: Internet computing) and supports information flows and processes over the Internet. Rather than a fat client, the user only needs a Web browser to access data and business logic capabilities on an Internet-connected server and database. Middleware applications emerged to manage the communication between operating systems and applications in a distributed computing system in a network. Distributed computing middleware can provide services for human interaction (that is, web requests) and machine interaction, including embedded systems in some industries.

Figure 7.2: Internet Computing

Hosted systems and the cloud

Deploying an enterprise or industrial application can be an expensive endeavor in terms of hardware and software costs, infrastructure, time, hiring and training, and maintenance and risk. On-premise systems require significant investment in time and budget.

The Internet made it possible for users to interact with a system from any Internet-connected device. Hosting services enable all or part of a system to be hosted and accessed from a remote location. Most hosting centers deploy applications in virtual servers using VMware, Oracle VirtualBox, Microsoft Hyper-V, or containers, and can spin-up the desired application on-demand. Containers are frequently deployed in hosted systems to deliver the operating system and other services. Container capacity can be scaled automatically, and since containers do not have a hypervisor, they can deliver better performance than virtualization.

Hosted services

Hosted services began as Application Service Providers (ASPs), which provided business applications in a hosted, managed, and centralized computing model. In the ASP model, the host provided and maintained a separate instance for each organization.

SaaS represents the evolution of the ASP model to a multi-tenant architecture. SaaS licenses software on a subscription basis and is centrally hosted, and is commonly used to provide office applications, database, development software, ERP, CRM, and other software.

PaaS provides a platform for the subscriber to develop and run applications on a hosted platform. PaaS systems can include a database, servers, storage, middleware, network, and other services necessary for application development and deployment.

IaaS supplies infrastructure resources on-demand from a centralized data center managed by the provi...