When you monitor data for over 5,000 assets across 50 customer locations and both those numbers keep growing year over year, you have be innovative. ABT Power Management (ABT) supplies and maintains industrial batteries and charging systems for both material handling equipment and DC back-up power for data center and telecommunication applications. ABT’s previous systems and processes were struggling to keep pace with their rapidly expanding business needs and were impeding their growth. As a result, it was difficult to manage the assets in a comprehensive, data driven and timely manner. ABT knew that as they continued to grow, to be able to maintain their customers’ experience of reliable uptime, lower operating costs, reduced capital spending, and operational power would require an innovative, leading edge, automated solution. Mariner was enlisted to help ABT design and implement Remote Automatic Asset Management System (RAAMS).
From IaaS to PaaS with Microsoft Azure
Originally designed in 2013, RAAMS was largely an Infrastructure as a Service (IaaS) solution – a classic “on premise” design, residing in the cloud. IaaS was the most appropriate architecture available at that time. Information from ABT’s customers’ batteries and chargers was collected in Microsoft SQL Server Express databases located on ABT’s customers’ sites or hosted by a third party vendor.
At RAAMS’s core were three Virtual Machines (VM). The first VM provided a Microsoft SQL Server database in which data was loaded via FTP and linked servers. Comma Separated Value (CSV) files were placed directly into SQL Server file tables via FTP and moved through Microsoft Service Broker queues to populate data tables while linked servers were used to populate these same tables directly through Data Manipulation Language (DML) statements. A second VM contained Microsoft SQL Server Integration Services (SSIS), SQL Server Analysis Services (SSAS), SQL Server Master Data Services (MDS), SQL Server Reporting Services (SSRS) and a data warehouse which was refreshed twice daily. A third VM ran Tableau Server to support advanced visualization and analytics. This architecture worked well, but had some limitations. On the data warehouse side it was restricted to two data refresh runs per day. On the data collection side specialized knowledge of SQL Server and a great deal of cooperation from individual customer’s network administrators was required, meaning ABT didn’t have the control over nor simplicity in monitoring their assets.
The Microsoft Azure platform continues to evolve rapidly and the current and planned Platform as a Service (PaaS) offerings now have significant advantages over the IaaS model in meeting ABT’s needs in terms of capabilities, features and cost. Replacing FTP and linked servers with Microsoft Event Hubs and Stream Analytics means the only requirement to monitor a remote site is an internet connection minimizing the required interaction by the customer or ABT personnel.
Elements of the Microsoft Azure IoT Suite
The release of the SQL Azure Data Warehouse allows real time reporting and decisions instead of data being refreshed only twice a day. SQL Azure Data Warehouse is also Polybase, which homogenizes access to SQL Azure databases, blob storage and Hadoop, handling all of the ways in which data may happen to arrive and removes limitations on the type and format of data that devices can report. A daily summary of work orders from a third party vendor can be written to blob storage and the steady stream of battery temperature telemetry can be placed in Hadoop throughout the day. In addition, PaaS works on a consumption based pricing model whereas the previous VMs were sized (and thereby priced) for the greatest anticipated workload. Moving to the consumption-based model yields significant cost savings.
As noted above, this new architecture leverages Microsoft Event Hubs and Stream Analytics, which are already available components of Microsoft’s Azure IoT Suite (IoT: Internet of Things) which is expected to be fully released later this year.
Subsequent articles will go into more depth about the Neuron Application, its interaction with Event Hubs and Stream Analytics and how it simplifies moving data from remote sites to the cloud. I’ll also delve into the SQL Azure Data Warehouse and how it can make real time data available to Power BI and Mobile and API Apps in Azure App Service to help drive decision making.
Learn more about Azure IoT and RAAMS
To learn more about ABT’s RAAMS and how it has positively affected their business, read the blog posts, Power Up with IoT from Microsoft’s IoT team and Predictive Maintenance with IoT in the Midmarket from Mariner.