WHITE PAPER ON Operational Analytics www.htcinc.com
Contents Introduction... 2 Industry 4.0 Standard... 3 Data Streams... 3 Big Data Age... 4 Analytics... 5 Operational Analytics... 6 IT Operations Analytics... 7 Conclusion... 8 Acronyms... 9 Reference... 9 1
Introduction In the current world, data is the asset of enterprises as they are used to generate knowledge and build decisions. It is termed as the raw material of 21st century. Data available in various forms is spawned by machines that are armed with electronic gadgets like motion sensors, cameras, microphones, GPS (Global Positioning System) and accelerometers, ATM (Automated Teller Machine) transactions, mobile phones, credit cards, and other electronic gadgets. Data is also produced by social network sources like Twitter, YouTube, Facebook, emails collections, web consumer clickstreams, mobile barcode readers, Radio Frequency Identification (RFID) sensors, and countless other sources. In today's fast moving, highly interconnected global business environment, enterprises are relying on complex systems built out of interconnected electronic sensors and gadgets built around Industrial Internet nicknamed as Industry 4.0 standard for their daily operations. These systems also called as Cyber-Physical Systems (CPS) are producing data with high volume, velocity, and variety for facing the competition. Generated data streams are varied and carry a lot of systems and business information that are critical for business success. High speed data in motion holds the key to valuable insights. Gaining visibility into this stream and the critical enterprise operations that they support can mean the difference between success and failure. To achieve a complete view of operational performance for enabling the rapid remediation of breakdowns, savvy enterprises are investing heavily in analytics capabilities. Since the insights are time sensitive, the business units have to act on those insights immediately. 2
Industry 4.0 Standard Data Streams The interconnected complex IT systems have become the prime movers of industrial production systems playing on economics around the world. Systems are built around machines, storage systems, and supplies that adhere to a defined standard and are linked up as cyber-physical systems (CPS). With industrial internet, also known as Industry 4.0, the systems have started to revolutionize the operating methods. The systems are responsible for digitization of inter and intra, horizontal and vertical value chains, and are directly responsible for the product and service portfolio delivery of enterprises. This new generation breed of systems has given us the ability to control granular data streams thereby opening up new world of possibilities leading to better insights. Streaming data is different from other kinds of data due to varying operational attributes which are attached to them. It is often loosely structured compared to other datasets. For example, the amount of email data streams being generated in an organizational context is quite high and fluctuates over time frame. The use of biometric devices across an organization convey a lot of insight on the organizational affinity and mobility. Data switching traffic sensors convey a lot of insight on the network traffic. The data is always available and new data is always being generated. Downtime for the primary collection system means that data is permanently lost. Insights are generated with a suitable mix of data streams and hence complexity is quite high. Future plants and systems will have clearly defined data connector interfaces for excellent data visibility. Emerging data churning technologies have the capability to flexibly replace machines along the value chain. Industry 4.0 emphasizes the idea of consistent digitization and linking of all productive units in an economy. Information on Event Processing, an emerging technology that helps to achieve actionable, situational knowledge from large scale event streams in real-time, is an interesting area for novel applications. Having carried out extensive research in the space of data streams, HTC's Research and Development team has generated white papers for the understanding of users at large. Refer the - Event Processing Systems the Way to Improved and Quicker Decision- Making white paper. 3
Big Data Age The physical and digital worlds are meeting at a very high speed. Most of the physical process are getting instrumented with sensors, telematics, and devices that drive ever-growing data. The major challenge that the modern world faces is the flooding digital data. Collecting, storing, and analyzing data from industrial sensors, network logs, and other machinery connected to the Industry 4.0 has become more feasible because of the emergence of big data technologies. The leading open-source software framework available for scalable, reliable, and distributed computing is Hadoop. It is changing not just the technology but also the economics of data storage and data warehousing. With solutions like Hadoop Distributed File System (HDFS), Hive, HBase, Pig, Oozie, Zoo-keeper, Flume, etc., Hadoop has become the low cost industry standard ecosystem for securely analyzing high volume data from a variety of enterprise sources. A simple overview of the ecosystem is shared here: 4
Analytics Arriving at a decision or gathering information for decision-making is not a simple and direct task. To derive information or gain knowledge with analysis, a suitable combination of aggregated data streams as raw material and analytic tools are required. Generally, analytics refers to analysis of data using Pareto analysis, trending, seasonality, regression, correlation, control charts, and other statistical techniques. It is not always a row of numbers noted from a gauge. Today, it is a combination of data streams that adds value to the organization. When analytics is integrated into business and decision-making processes the insight flows automatically to thousands of knowledge workers, and thousands of decisions made each day by people or computers. Big data analysis requires interconnected set of solutions from acquiring the data to making decisions for repeated analysis. Organization's early use of analytics focused on developing data dashboards that depict information in graphical form. This approach makes it easy for executives to detect trends and take action, and it was an important step in turning information into insights more quickly. According to Wikipedia, analytics is the process of developing optimal or realistic decision recommendations based on insights derived through the application of statistical models and analysis, against existing and/or simulated future data. When analytics is applied for the day-to-day operations, one gets into operational analytics. The operational transactions have decisions with it and each action is driven by a decision. The goal of any analytic solution is to provide the organization with actionable insights for smarter decisions and better business outcomes. Different types of analytics, however, provide different types of insights. Analytics solutions are of three principal types: Ÿ Descriptive - uses business intelligence and data mining for analyzing: What has happened? Ÿ Predictive - uses statistical models and forecasts for analyzing: What could happen? Ÿ Prescriptive - uses optimization and simulation for analyzing: What should we do? 5
Operational Analytics Big data and predictive analytics are used by business and industries to change and improve the processes and operations majorly. Industry 4.0 interconnects the various production level sensors in the value chain. Operational analytics enable the businesses to analyze and mine stream the machine generated sensor data and / or production data to provide real-time insight into enterprise operations. Data streaming enables real-time analysis to take place. Instead of querying static data, real-time data streams are continuously evaluated by static questions. When data is analyzed in real-time it produces numerous benefits and the details are discussed in the Managing Realtime Data Streaming Effectively with CEP Solutions white paper. Operational Analytics automates analytics for helping end users (or systems) during the decision-making process itself, leading to Operational Intelligence. Decisions, whether tactical or strategic, are critical to the success of every organization says Davenport, in his recent paper, How Organizations Make Better Decisions, (Jan 2010). The basis of operational analytics built over Industry 4.0 standard is the increased availability, and integrated use of relevant data by connecting all products, resources, and enterprises involved in the value chain. It includes the ability to generate additional value from available data and for maximizing customer benefits. This requires a fundamental transformation of processes, the product, and service portfolio as well as the existing business models. The granular level visibility into systems and devices with analytics and insights help enterprises to increase their operational efficiencies and reduce costs. 6
IT Operations Analytics IT Operations Analytics (ITOA) is also known as Advanced Operational Analytics or IT Data Analytics. It encapsulates technologies that are primarily used to discover complex patterns in high volumes of 'noisy' IT system availability and performance data. There are varieties of analytic dashboards: Ÿ Operational Dashboards: Helps the front-line workers and supervisors to track the core operational processes Ÿ Tactical Dashboards: Enables managers / analysts to track / analyze departmental activities, processes, and projects. Ÿ Strategic Dashboards: Allows executives and staffs to track the progress in attaining their strategic goals Operations Dashboard is a multilayered application built on business intelligence and data integration infrastructure, enabling enterprises to measure, monitor, and manage business performance more effectively. They focus on monitoring functions more than analysis or management functions. Tactical Dashboards emphasize analytical functionality more than monitoring or management features. Analytical functionality enables users to investigate the root causes of problems, issues, or trends. Strategic Dashboards emphasize management features more than monitoring or analytical functionality. Operational Analytics applies to organizations of all sizes across industries along the entire length of supply chain. Understanding customers' behavior, enhancing their experience and profitability is equally important regardless of the customer base size. In Operational Analytics, numerous combination of transactional data, such as product aggregations or aggregations of customers who are geographically close may be performed. Descriptive models are used to determine demand forecast, manufacturing and distribution costs and cost relationships, future raw material costs and availabilities, and a variety of other parameters and relationships required by the Operational decision database. 7
IT Operations Analytics Cont. Conclusion HTC supported a renowned data storage company for designing their operational dashboard to monitor the production system. Production level details were scattered across the spectrum. Such details were identified and cataloged. Based on the organizational needs, suitable metrics were identified and mapped to the organizational goals with relevant risk measures. As planned, information from multiple systems were extracted for further analysis. The identified matrices were mapped to the operations dashboard. Suitable triggers were planned and run for the benefit of end users leading to good visibility. Operational Analytics uses big data technologies for the latest applications to analyze machine data and gain insight, which gives better business results. The data generated by machines collected by the IT systems contain valuable insights. IT Operational Analytics automates the process of collecting and organizing data for locating patterns that help in identifying business results and improving system performance. In the current market, there are several software solutions for operational analytics that help in easily transforming the disconnected data (that reside in disparate systems) to actionable information. 8
Acronyms The acronyms used in this white paper and their expansion is provided below: Acronym ATM CPS GPS HDFS ITOA RFID Expansion Automated Teller Machine Cyber Physical Systems Global Positioning System Hadoop Distributed file system IT Operations Analytics Radio Frequency Identification Reference 1. Operational Analytics Cloudera: https://www.cloudera.com/content/dam/cloudera/resources/pdf/whitepaper/ WP-Operational-Analytics-101.pdf 2. Ten IT-enabled business trends for the decade ahead: http://www.mckinsey.com/industries/high-tech/our-insights/ten-it-enabledbusiness-trends-for-the-decade-ahead 3. Think Act: https://www.rolandberger.com/media/pdf/roland_berger_tab_industry_4_0_2 0140403.pdf 4. Descriptive, Predictive, and Prescriptive Analytics Explained - The two-minute guide to understanding and selecting the right analytics: https://halobi.com/2014/10/descriptive-predictive-and-prescriptive-analyticsexplained/ 5. Using Performance Analytics Dashboard: http://wiki.servicenow.com/index.php?title=using_performance_analytics_das hboards#gsc.tab=0 9