Fog-Computing-Virtualizing-Industry-White-Paper
2 likes358 views
The document discusses how fog computing and distributed intelligence platforms can disrupt the energy sector by enabling a more efficient and reliable industrial network. It describes how fog computing involves performing compute, storage, control and networking functions closer to industrial devices and sensors rather than relying solely on remote cloud services. This allows for lower latency responses, better security and a more adaptable system. The document provides examples of how fog computing could improve microgrids, SCADA integration, grid balancing and create a distributed system platform in the energy industry.
Fog-Computing-Virtualizing-Industry-White-Paper
- 1. Fog Computing – Virtualizing Industry How Distributed Intelligence & Low Cost Industrial Fog Computing Will Disrupt Energy Written by: Graham Beauregard, CTO, LocalGrid Technologies Inc. March 26, 2015
- 2. © 2015 LOCALGRID TECHNOLOGIES INC. | FOG COMPUTING – VIRTUALIZING INDUSTRY 2 Executive Summary The Industrial Internet of Things is driving a fundamental transformation of industrial networks and systems with solutions that will impact how people interact with machines. The connection of people, machines, sensors, and a large interconnected network, combined with lower cost industrial compute power, will disrupt existing business models and fundamentally change the competitive nature of many industries. Nowhere is this more evident than in the power sector, where rapid decline in the cost of renewable energy, policy change, and the adoption of smart, distributed, and connected assets, is driving a revolution in how power is generated, consumed and managed. Investments to leverage Big Data will improve insight and drive efficiency; however, perhaps more than any other industry, the power grid must be adaptive, responding quickly to disturbance, and above all, be reliable. There are challenges with Big Data solutions when considering the bandwidth, latency, and security of transmitting vast amounts of data from industrial systems to the Cloud. Many questions remain around this transformation, including how it will impact existing infrastructures, value chains, business models, and workforces. This paper provides insight into the strategies industrial sectors employ to improve the efficiency and reliability of their industrial networks and systems. It details emerging Fog Computing technology and how efforts to decentralize systems and architectures can compliment Big Data solutions to overcome many of the existing business and technological barriers to adoption. The paper explains how industries will utilize distributed intelligence platforms to manage connected devices, sensors, and machines to create a more secure and efficient network that has the flexibility to evolve with changing requirements and needs.
- 3. © 2015 LOCALGRID TECHNOLOGIES INC. | FOG COMPUTING – VIRTUALIZING INDUSTRY 3 Table of Contents Executive Summary ............................................................................................................ 2 Table of Contents ............................................................................................................... 3 Background ........................................................................................................................ 4 Industry Trends .................................................................................................................. 5 The Industrial Fog ............................................................................................................... 6 LocalGrid Fog Computing PlatformTM ................................................................................. 8 Virtualization of Function …………………………………………………………………………………………..9 Security ......................................................................................................................... 10 New Application Development .................................................................................... 11 Sample Applications ......................................................................................................... 12 MicroGrids .................................................................................................................... 12 SCADA Integration ........................................................................................................ 13 Grid Balancing .............................................................................................................. 14 Distributed System Platform ........................................................................................ 15 Conclusion ........................................................................................................................ 15
- 4. © 2015 LOCALGRID TECHNOLOGIES INC. | FOG COMPUTING – VIRTUALIZING INDUSTRY 4 Background The Industrial Internet, or the Industrial Internet of Things, is an emerging set of technologies and standards for the interconnection of industrial devices and networks. The impact these technologies and standards will have on industry will be of the same scale that the growth of the Internet has had on business. Some leading experts have estimated that the impact may be the greatest change to how industry operates since the industrial revolution. The Industrial Internet is the connection of people, data and intelligent machines. It enables a higher degree of decision, support and a shortened timeframe for opportunity recognition, enhancing growth potential, and competitiveness. Big Data and the value of the connected network are the foundations driving this change. In a recent survey of multiple industries completed by Accenture and GE, 80 to 90 percent of responding
- 5. © 2015 LOCALGRID TECHNOLOGIES INC. | FOG COMPUTING – VIRTUALIZING INDUSTRY 5 companies indicated that Big Data analytics is either the top priority for the company or in the top three1 . If companies do not implement these solutions they will lose market share, and become less competitive. Metcalfe’s law states that the more machines (end-‐points) connected together in an interoperable network, the more value can be derived from the network2 . Modernizing the North American electrical grid, for example, will enable distribution utilities to become smarter and more efficient in the way energy is produced and consumed. The Industrial Internet will break down previously siloed, distinct systems and networks and evolve them into one holistic, distributed and integrated platform that supports the function of many applications. The Industrial Internet is bringing together software technologies and traditional machines, devices, and sensors to build intelligence into the networks, while this provides many benefits; it also creates challenges of how to make sense of all the intelligence (data) being collected. Utilities report that there are too many moving parts of their system to allow them to readily associate the cause and effect when problems arise. Although having software intelligence at all endpoints -‐ such as traditional energy sources, distributed renewable energies, distributed energy storage, and electric vehicle loads -‐ enables a smarter and more granular scope of the electrical grid, utilities are still looking for strategies to filter data and enable real-‐time decision making. Industry Trends The emergence of Big Data, gathering information from machines, sensors and devices within an industrial network or system will create new high-‐volume, real-‐time data streams. This influx of data can be overwhelming for many existing systems and the complexity of the data may be hard to consume with the existing infrastructure. Current systems are typically centralized, transmitting all data to a central server for processing and analysis. But what if the processing and analysis could be performed on the machine, sensor, or device itself and in turn only transmit data that meets predefined parameters? This type of distributed intelligence network is evolving as part of the Industrial Internet, particularly in the energy industry.
- 6. © 2015 LOCALGRID TECHNOLOGIES INC. | FOG COMPUTING – VIRTUALIZING INDUSTRY 6 Figure 1: Comparison of North American utilities to the number of electric endpoints GTM Research released a report, Utility Smart Grid Outlook in North America 2013, profiling 40 North American utilities with representation of 46.6 million electric endpoints on the current power grid3 , including 29.2 million smart meters deployed, accounting for only 38% of planned smart meter deployments at the time (see Figure 1). With the onslaught of new endpoints (or grid edge devices) in the form of distributed energy devices -‐ such as photo-‐voltaic solar, wind turbines, energy storage and electric vehicle loads -‐ it is going to be increasingly complex to manage all of these assets and even more important to have insight into their real-‐time operation. Distributed intelligence, or intelligence at the edge, is gaining acceptance in the energy industry as a necessary evolution to the existing grid, but the concept can be applied across industry sectors. Placing intelligence in edge devices provides better insight to where failures and outages are happening in real-‐time and allows grid operators to use better information to gain a much deeper insight into network operation. The Industrial Fog Industrial monitoring and control systems operate locally (on the same network). These systems interface with sensors, machines, and devices. This has been, and continues to be required as control latency and jitter must be kept low, reliability must be high, security is critical, and massive data volumes are generated in order to satisfy operational requirements. In industrial applications, system failure can be catastrophic
- 7. © 2015 LOCALGRID TECHNOLOGIES INC. | FOG COMPUTING – VIRTUALIZING INDUSTRY 7 and connections to Big Data solutions in the Cloud cannot replace the need for local action. Today, many devices and systems are being connected to the Private or Public Clouds in order to centralize data aggregation, perform large scale analytics, allow for extreme scalability, and provide wide-‐scale access of information to operators, business systems, and customers. This technology concept is better known as the Industrial Internet of Things, or the IIoT. Traditional cloud and machine-‐to-‐machine (M2M) platform solutions rely exclusively on the Internet for all user and device communications. In these scenarios, we are moving data from devices into the Cloud and then onward to applications for processing. While this does provide great benefit in terms of scalability, maintainability, and wide-‐scale access to data for users, it unfortunately does not satisfy the requirement for low latency, high reliability, security, and management of massive volumes of data demanded by industrial systems. In order to realize all of the benefits of the cloud and satisfy the requirements of industrial monitoring and control systems, LocalGrid has designed an architecture that combines both decentralized distributed intelligence with interfaces to Cloud and server back-‐end systems. This area of decentralized distributed intelligence is known as the “Fog” and allows for local analytics, control, and direct device-‐to-‐device communication, avoiding the need for a connection back to the Cloud in order to operate. This is a paradigm shift in which we are now moving applications from the Cloud closer to the data at the edge. Combining the power and flexibility of the Cloud with the benefits of the Fog provides industrial customers the best of both worlds – satisfying the latency and control requirements of critical systems, while delivering the benefits of Cloud computing. Within the Fog, devices communicate with each other peer-‐to-‐peer requiring no intermediary broker or servers (see Figure 2). This creates architectures with no single point of failure and highly resilient systems that tolerate partial system degradation without complete system collapse. Peer-‐to-‐peer communication has much lower latency, higher performance, and determinism – critical features in industrial systems. Applications live in the Fog, where data can be analyzed, stored, and distributed locally. Only important extracted information is passed upwards to the Cloud for long-‐term storage and post-‐processing. This greatly reduces the amount of bandwidth and
- 8. © 2015 LOCALGRID TECHNOLOGIES INC. | FOG COMPUTING – VIRTUALIZING INDUSTRY 8 centralized storage required, inherently reducing cost over time. Smaller subsets of devices may operate in device clusters within the Fog, keeping data co-‐located with devices that acquire and require it. This reduces the potential for security attacks, which would normally compromise the entire network, to a single subset of the network. Figure 2: Fog and Cloud Architecture With the continued rise of processing power of embedded devices alongside the declining cost of compute, we are entering a period in which Fog architectures are a reality and will ultimately become a necessity. The key to managing the vast amount of devices and data is through a massive decentralization of intelligence, or distributed intelligence. Distributed intelligence is the premise of Fog networks, and is the next generation of computing and information technology. LocalGrid Fog Computing PlatformTM LocalGrid Fog Computing PlatformTM is an essential component to deliver Industrial Internet of Things architectures. Given the importance and respective complexity of Fog architectures, reducing this complexity and therefore reducing risk is critical. LocalGrid Fog Computing Platform is a standards-‐based software framework where complex in-‐
- 9. © 2015 LOCALGRID TECHNOLOGIES INC. | FOG COMPUTING – VIRTUALIZING INDUSTRY 9 field network architectures that include analytics, monitoring, control, application deployment, and security, can be implemented. Building systems on standards-‐based off the shelf platforms allows end-‐users to focus on their unique applications without worrying about the complexity of Fog networks. Specific Benefits of the Fog Computing Platform include: § Peer-‐to-‐peer device communication, § Low latency communication – on the orders of milliseconds and microseconds, § Low jitter communication, § No single point of failure, § Inherent redundancy of applications, § Communication protocol conversion using a common data model and interface, § Application deployment and device virtualization, § Inherent security through encryption, authentication, and non-‐repudiation, § Simplified application development process, § SDK available in C, C++, C#, Python, and LabVIEW, § And more… Virtualization of Function With the rise in available compute power on embedded devices, there is an opportunity to push additional intelligence and applications to the edge of the network. The capability of running applications at the edge allows a new degree of virtualization to occur where a collection of virtual software based “devices” can all run within a single piece of hardware. The same way virtualization has revolutionized the way Cloud scale servers operate, Fog software platforms converge many functions into a single device that will contain multiple personalities/ functionalities executing in parallel. LocalGrid Fog Computing Platform enables many virtualized devices to co-‐exist on numerous supported hardware targets. Multi-‐target support, and virtualization of functions, can reduce operational complexity, extend the life of existing assets, and
- 10. © 2015 LOCALGRID TECHNOLOGIES INC. | FOG COMPUTING – VIRTUALIZING INDUSTRY 10 ensure efficient use of compute resources. Applications for analytics, control, or implementing virtual devices can be developed using the provided LocalGrid Fog Computing Platform cross-‐language SDK, and then deployed and managed through provided tools enabling complete life-‐cycle management of industrial applications running in the field. In evolving industrial systems, legacy equipment must co-‐exist with modern equipment to reduce costs and extend asset life. Although it is not practical to tear-‐down and replace all legacy equipment that may still have significant functional life, to fully benefit from modern technologies and communications protocols, devices must be able to speak to one another. LocalGrid Fog Computing Platform performs protocol normalization in the field, adapting legacy and modern protocols to a common open standards-‐based protocol to allow devices to speak naturally to each other at the edge of the network. Performing protocol conversion at the edge creates new communication paths that previously could not exist, and does not require the tear-‐down of pre-‐existing communication paths, instead it runs in parallel to them. This allows for multi-‐vendor and multi-‐protocol interoperability while allowing legacy communication paths to be maintained – enabling a natural evolution of systems. Systems may be modeled, deployed, managed, and secured through LocalGrid’s configuration and management tools and APIs. Security The widespread deployment of sensors, devices, and other things creates a natural tendency for industrial networks to sprawl, extending the boundary of security. In many cases, this network sprawl includes areas where network owners may not have control over physical access, such as a customer premises or public site. Security at the Fog level must be carefully considered, as it is the last line of defense before connection to the Cloud where access to vast amounts of data is possible. LocalGrid Fog Computing Platform implements an open standard based security approach that was designed specifically for the needs of high performance distributed assets to communicate peer-‐ to-‐peer. Our solution: § Provides authentication, authorization, non-‐repudiation, confidentiality and integrity of data, § Defends against unauthorized access, tampering and replay, § Operates without centralized servers for high performance, scalability and availability, and § Integrates with existing security infrastructures and hardware acceleration
- 11. © 2015 LOCALGRID TECHNOLOGIES INC. | FOG COMPUTING – VIRTUALIZING INDUSTRY 11 The security model is completely decentralized, and does not rely on connection to a server or the Cloud to operate once provisioned and does not sacrifice performance for security. Different industries and customers require different levels of security, which is why we have included a security model which is based on a plug-‐in approach, allowing for industry specific or regulatory specific encryption and cryptography technologies to be added as required by end-‐users. The platform does not prescribe which security technology you must use, but does include a default set of technologies that is more than sufficient for most use cases. To application developers, this security model is transparent, adding very little complexity and effort to deploy highly secured Fog based architectures. New Application Development LocalGrid vRTUTM (Virtual Remote Terminal Unit) is a LocalGrid software product that transforms any intelligent automation controller, router, or gateway into a fully featured RTU without the additional hardware normally required. It collects and aggregates signals from remote I/O devices and sensors, enables bi-‐directional interoperability across interfaces and protocols, provides remote control capabilities, and integrates with legacy SCADA systems. By transforming edge inter-‐device communications and protocols into compatible open standards based outputs such as Modbus, DNP3, IEC 61850, DDS, and MQTT these protocols become consumable across the network by all devices and systems in a secure manner. LocalGrid PQATM (Power Quality Analyzer) is a powerful solution that integrates distributed data architectures with advanced real-‐time and historical power analytics. By deploying a software-‐based solution on a flexible hardware platform, utilities are able to use an extensive array of standard power quality data (e.g. harmonics, phase, overvoltage detection, etc.) in an open-‐architecture platform that allows the solution to evolve as requirements change, without incurring new hardware or installations costs. LocalGrid Fog Computing Platform provides the standard framework for these and other applications to work together, unifying communication standards between edge devices and simplifying application development across hardware platforms.
- 12. © 2015 LOCALGRID TECHNOLOGIES INC. | FOG COMPUTING – VIRTUALIZING INDUSTRY 12 Sample Applications MicroGrids LocalGrid Fog Computing Platform is especially well suited for modern day distributed smart grid architectures, such as MicroGrids. MicroGrids are inherently decentralized, their purpose is to create a locally managed and self-‐sustaining subsection of the traditional power grid in the event that a power outage occurs. MicroGrids can be used to manage hospitals, sporting arenas, military bases, major events centers, and clusters of homes. In order to accomplish this, MicroGrids must contain not only significant loads to manage, but significant amounts of distributed generation, such as solar and wind, and energy storage. LocalGrid Fog Computing Platform integrates data from all of the sensors, devices, and machines that constitute the MicroGrid and allows for localized control of the assets and loads – delivering a high level of resiliency and support for continued operation under degraded conditions. An outline of the software structure is shown in Figure 3. Figure 3: MicroGrid Architecture Fog architecture pushes intelligence to the edge of the network and decentralizes decision-‐making and analytics. In the past, energy management and grid automation control was performed through centralized analytics and control. A centralized control system has several downsides in a MicroGrid environment: § A failure of the central controller can disrupt the entire MicroGrid, § Expensive hardware is required for the central controller,
- 13. © 2015 LOCALGRID TECHNOLOGIES INC. | FOG COMPUTING – VIRTUALIZING INDUSTRY 13 § System maintenance requires complete system shutdowns, § Scalability and expansion is a complex and expensive task, § Security attacks on a central controller can take down the entire network, § Central management works against the nature of a MicroGrid, which is intentionally distributed. In a modern grid (Figure 4), it is typical to find a variety of hardware both multi-‐vendor, and varied in age and performance characteristics. Critical to constantly evolving technology and rapidly growing infrastructure is the interoperability of systems. LocalGrid Fog Computing Platform supports direct edge-‐to-‐edge protocol conversion to ensure interoperability across all hardware and devices in the grid or MicroGrid. This multi-‐protocol support ensures a simple evolution path for grid operators and reduces the risk of non-‐interoperable module rollouts of new technology and increasing levels of distributed generation. The platform fully supports conversion to and from the following industry standard communication protocols: Modbus serial, Modbus TCP, DNP3 serial, DNP3 TCP, IEC 61850, and others. Figure 4: Modern Smart Grid A distributed MicroGrid monitoring and control architecture provides the highest reliability and safest solution while also providing high levels of scalability and flexibility in terms of system rollout and maintenance. SCADA Integration The Smart Grid is transforming utility operations and pushing IT (Information Technology) across its traditional boundary into OT (Operational Technology), merging the two previously distinct categories for smarter, more cost effective, and more reliable operation. However, some significant barriers to this effort exist. Legacy devices
- 14. © 2015 LOCALGRID TECHNOLOGIES INC. | FOG COMPUTING – VIRTUALIZING INDUSTRY 14 deployed to the field may not support the protocols or functions required. System and data models must be rationalized across many existing systems. Bandwidth and integration costs can be prohibitive, especially when the full value of the new effort is not proven. One solution gaining acceptance in the industry is the use of distributed intelligence. This includes development of a field message bus supporting a unified data and system model and peer-‐to-‐peer communication4 . LocalGrid Fog Computing Platform supports edge protocol conversion to integrate legacy devices, a rational data and system model to eliminate silos between systems, and integration in the field with standard interfaces to back-‐end systems. This in-‐field integration can reduce bandwidth costs (by pre-‐ processing data in the field) and reduce integration costs by eliminating the need to modify existing IT and OT systems. Intelligent edge devices can serve as a gateway, aggregator, protocol converter and an application platform, all in one device, such as a smart network switch or automation controller. Device virtualization provides a cost effective migration path to integrate new functions and legacy equipment. In virtual IEDs (Intelligent Electronic Devices) that are based on generic hardware platforms, the task of each component can be performed simultaneously on the hardware. LocalGrid Fog Computing Platform supports many intelligent devices and compute platforms and provides an application environment for data sharing across these virtual applications. Grid Balancing Grid balancing is the critical process of matching power generation with power consumed, traditionally achieved by ramping up or down power output to support current load. This process becomes more complex when there are intermittent or unpredictable power generation sources, such as renewable solar and wind, which are most efficient when allowed to generate at maximum output when resources are available (when the wind blows or the sun shines). The current grid is designed assuming generation will follow load, but to support greater amounts of renewable generation the grid must change, load must follow generation 5 . To make this
- 15. © 2015 LOCALGRID TECHNOLOGIES INC. | FOG COMPUTING – VIRTUALIZING INDUSTRY 15 fundamental change to the grid, utilities will use smart infrastructure, various forms of MicroGrids, and distributed, managed loads to manage and shift demand. LocalGrid Fog Computing Platform provides fundamental building blocks to deploy these new technologies efficiently to the grid. A decentralized approach to managing these new systems will drive efficiency and lower overall costs but requires a new IT framework built on distributed intelligence, low-‐latency communications, and autonomous actions. Distributed System Platform The emergence of next-‐generation technologies is creating a fundamental shift in the utility business model and new regulations are emerging to address these changes and create a distributed, consumer-‐focused energy system. New York State Energy Research and Development Authority (NYSERDA), with their Reforming the Energy Vision (REV) initiative, are leading one of the most forward-‐ looking transformation initiatives. Under REV, electric utilities will evolve to act as Distributed System Platform Provider’s (DSPPs) who actively manage and coordinate distributed resources, providing customers with improved electricity pricing structure and creating new markets6 . The platform will provide safe, reliable, and efficient electric services by integrating diverse energy resources to meet customers’ and society’s evolving needs. This type of approach will expand the value of the electric system and provides a mechanism to enhance economic and environment value through a fully integrated grid. For this new operating model to be fully realized, a common standardized IT/OT platform, like LocalGrid Fog Computing Platform, is required by the DSPP. This platform must support interoperability across hardware and systems to connect all of the assets and devices into one unified model, while also encouraging flexibility and scalability to evolve the grid as needed. Conclusion The current industrial landscape is being challenged like never before. The impact of the connected world is really just starting to impact traditional industries. As systems become vastly distributed on a global scale and the reliability of these assets becomes critical to our safety and security, new technologies that support interoperability between previously distinct systems will become the standard.
- 16. © 2015 LOCALGRID TECHNOLOGIES INC. | FOG COMPUTING – VIRTUALIZING INDUSTRY 16 Forward-‐thinking utilities are striving to effectively deploy distributed energy resources to the traditional grid, deploying distributed intelligence solutions will enable them to support smart technologies and increase the reliability of the grid. In energy and other industries, more intelligence will be deployed nearer to assets, leveraging inexpensive processing power on existing devices, machines and sensors. Interconnections between these devices and systems lead to an increase in the value of the entire network through an unprecedented sharing of data between systems and devices. Distributed intelligence, like LocalGrid Fog Computing Platform, will provide the foundation for industry to build these networks. As industrial systems become more complex and globally distributed, the ever-‐increasing processing power of embedded devices, alongside the increasing value of interconnected systems, will make Fog architectures a necessity. 1 GE and Accenture, Industrial Internet Insights Report for 2015, 2014. 2 Metcalfe’s law states that the value of a telecommunications network is proportional to the square of the number of connected users of the system. 3 Emma Ritch, GTM Research, Utility Smart Grid Outlook in North America 2013: Technologies, Strategies & Case Studies, April 2013. 4 The Smart Grid Interoperability Panel (SGIP), in partnership with industry, is creating an Open Field Message Bus (OpenFMB) framework, www.sgip.org 5 A recent ARPA-‐E funding opportunity from U.S. Department of Energy (DE-‐FOA-‐0001289: NETWORK OPTIMIZED DISTRIBUTED ENERGY SYSTEMS (NODES) has estimated the value of replacing 4.5GW of spinning reserves with flexible loads and distributed energy resources as a 3.3 Billion/year USD opportunity in the PJM Market. 6 Reforming the Energy Vision, NYS Department of Public Service Staff Report and Proposal, Case 14-‐M-‐0101, April 24, 2014.