April 04, 2016
Designing an IIoT Solution for Public Cloud
By: Guanyi Sun, Andy Guo Ma, & Bin Xie
The Industrial Internet of Things (IIoT) is more valuable than the sum of its parts. It is not enough to be “connected.” The goal is to be pervasive, transformative and intelligent

Building on this idea, the Accenture Technology Labs IIoT R&D group is combining scalable computing and storage bandwidth with advanced techniques in cloud platform to design and build a public cloud-based IIoT solution. Moving an IIoT solution to the cloud can bring tremendous productivity for companies participating in an ecosystem. This blog talks about the design of an efficient IIoT architecture and the benefits of implementing this solution on different public cloud platforms. 

Tech Labs is partnering with the Accenture Utilities industry group on a solution to help a utility client upgrade an existing system to make its business smarter and more efficient. With the company’s existing system, it takes 30 to 45 minutes for the data from an IIoT sensor located on water meter to be converted into a visualization that the utility can use to make a decision. The client’s new solution needed to introduce more feature data, monitor equipment status and calculate predictive maintenance results—all with a processing time of mere seconds. At the same time, the client needed to reduce maintenance costs and handle future requirements quickly and efficiently.

To address these varied requirements, Accenture architected a comprehensive IIoT solution with seven stages: ingest, process, store, analyze, enable, archive and access. As shown in Figure 1, the ingest stage collects outside data from different sources such as meters, sensors, handheld devices and even paper-based logs. After data cleansing and quick initiative analysis in the process stage, the cleaned data and first-step results are stored in different databases for various data structures (store stage). Later the enable stage drives the analyze stage to conduct deeper and more complex data mining for different business requirements. The utility achieves the needed data and analysis result through a visualization interface in the access stage. Finally, the data is stored in the archive stage.

Figure 1: Ilot system architecture

Figure 1: IIoT system architecture

Moving this IIoT system architecture to a cloud platform provides unlimited and scalable resources more cost effectively, while reducing the utility’s non-core tasks such as daily maintenance or software environment setup. In order to take this step, Accenture tested three public cloud platforms—AWS from Amazon, Azure from Microsoft and AliYun from Alibaba—to evaluate the differences caused by unique featured services on each public cloud. Table 1 lists each key service used.

Table1. Unique services in each cloud platform











Load Balance







Elastic MapReduce





Elastic Beanstalk







Azure ML


Azure Warehouse


Azure Storage

Load Balancer


Message Notify Service



Open Table Service


Open Data Process Service


Data Process Center

 Elastic Compute Service

 Object Storage Service

Server Load Balancer


Through this work, our IIoT R&D group successfully demonstrated how to establish an IIoT solution architecture based solely in a mainstream public cloud service, as well as how to transfer the solution among different cloud platforms.

Each cloud vendor offers compelling differentiation based on its featured services. For example, AWS provides good compatibility with mainstream open source software. Azure integrates with Microsoft Office, which offers a familiar user interface. And AliYun excels at big data analysis, especially at the terabyte and petabyte level.

Accenture Labs will continue monitoring the unique services of these and other public cloud vendors. We are also working on defining benchmarks and scenarios to evaluate the performance of key services in each cloud platform. 

Popular Tags

    More blogs on this topic