See all apps

Remote Process Control using Reliable Communication Protocol

Submitted on April 03, 2013

What problem are you intending to solve?

We will remotely observe and control advanced manufacturing processes using software defined networking and gigabit broadband with high reliability, which is not possible with the TCP/IP Internet.

What is the technological approach, or development roadmap?

Today’s TCP/IP Internet is poorly suited to supporting an increasingly important range of time-sensitive applications because it builds a single path to deliver data between communicating systems. If data delivery is delayed or blocked due to path congestion (overload) or outright failure, the TCP protocol will find a better path, eventually, but the quality or even success of the application may have already been compromised. Furthermore, today’s Internet cannot make guarantees to applications that require reliable communication.

We will use Software Defined Networking (SDN) to deploy a protocol that can create and actively manage redundant paths between communicating devices and that will proactively build new paths before network congestion and faults degrade application performance. With replication of data packets and monitoring of multiple path performance, Internet communication will move beyond “best effort” packet forwarding of TCP/IP to reliable packet delivery. Thus, the Reliable Communication Protocol (RCP) project will create a fundamentally new Internet capability.

Combining redundant paths with gigabit broadband will create an ideal platform on which to build any number of time- and reliability-sensitive applications for which today’s Internet is insufficient or incapable. While our initial application focus is to support advanced manufacturing processes, there is clear potential to support applications in many other areas.

Our protocol concept is illustrated in the attached figure, which should be paired with the following caption.

Figure 1 caption: The communicating end systems are each connected to the Internet via a redundancy controller (RC). Using OpenFlow software defined networking the sending RC establishes redundant paths (green and blue), replicates packets, and sends data by both paths. The receiving RC notes the arrival of packets, forwards a copy of each packet to the end system, and matches replica packets. In matching replica packets the receiving RC can perform ongoing monitoring of path performance throughout the duration of the communication. Congestion or failure within a path (shown as a red link within the blue path) is detected by the receiving RC as delayed or failed ability to match a packet with its replica. Delay or outright failure on one path does not prevent the delivery of packets to the end system, thus communication is reliable. Delayed reception of a replica packet can trigger the reliable communication protocol to build a new path (purple) to restore path redundancy and maintain a desired level of reliability.

Consider an advanced manufacturing application using the reliable communication protocol to control the process taking place within the build volume of a remote 3-D printer. With a reliable connection a part designer could do much more than simply download from her computer to the printer the 3-D item shape and material(s) selection command sequence. Instead, her computer could send printer commands to execute a customized fabrication path and sintering energy to use within the build volume, monitor the part as it is forming, compare those observations with the design, and apply a control algorithm to adjust future printer commands to keep the part within tolerances, achieving a feedback control loop across the internet. The end goal could be to realize tighter control of process, materials, and tolerances, or it could be to reserve some portion of the printer command generation to a computer secured by the designer, thus concealing a portion of her manufacturing knowledge from the 3-D printing location. Generalizing this idea, one can imagine the designer establishing a connection to a third-party control algorithm service provider that can deliver advanced process knowledge or superior computing capability or both and so on. Reliable communication allows the components of the process and the associated information to be located where desired.

How will end users interact with it, and how will they benefit?

Typically, end users will employ applications that interact with our reliable communication protocol at a high level. In many cases we can imagine that there would be no need for the user to be specifically aware of using our protocol, not to mention having to interact with with it. A particular internet device or service would be programmed to establish a reliable connection and if it could establish a connection of sufficient quality might not even notify the user. For more demanding applications, the application or device might notify the user of the quality of the connection relative to that desired and allow the user to decide whether to accept or reject the connection. Being able to reliably control a manufacturing process that is remote to the user will empower engineers, artists, and experimenters who are not near the means of production. Manufacturing teams can each bring their specialized skills and capabilities together without regard to limitations of geography. Healthcare providers could reliably deliver remote diagnosis, advice, or even treatment, yielding better outcomes to patients who are far from an appropriate specialist or medical facility. For example, custom medicine tablets could be 3D printed in the home to the exact need and timing for the patient, with the printing (manufacturing) process monitored in sufficient detail and with high, reliable data rates to validate that the printer achieved FDA good manufacturing practice and, thus, a quality tablet. Another application would be to allow sharing of specialized manufacturing equipment more easily. For example, reliable internet communication may enable sufficiently high quality telepresense that a university laboratory could comfortably share its instruments with the wider community, bringing valuable capabilities to the off campus community and subsidizing the cost of the equipment, benefitting the laboratory.

How will your app leverage the 1Gbps, sliceable and deeply programmable network?

We will use the resources of the GENI Project to support the development of our protocol and demonstrations of its application. Team members Adams and Geske are Project Leads with GENI and can commit GENI resources to our experiments and invite others to work with us on these experiments. GENI provides direct assess to software defined networking in the form of OpenFlow networking hardware. The ability to build a fault-tolerant, fixed latency communication path does not exist with todays TCP/IP Internet, but OpenFlow switches can establish redundant communication paths and manage them. See U.S. Patent 4,523,273 [Adams and Siegel] for the networking idea that inspired this project. Delivering reliable communication is a complex task. Our roadmap for development envisages two areas of development and concurrent experimentation. First, using the GENI testbed we are setting up an OpenFlow slice having two nodes (initially) with OnTimeMeasure Node Beacons, between which we can install and switch flows using two different paths to demonstrate an application for remote control of a manufacturing process. The command line tools in OnTimeMeasure to query the delay and loss measurements from Root Beacon and display the health of the two paths on a web page. Second, we are coding a preliminary version of the protocol running in a virtual environment on computers in a single lab. On this implementation platform, and as it is developed towards a beta readiness, we will conduct experiments on the performance and behavior of the protocol. Protocol experiments will be carried out by team members at all our sites: Purdue, Ohio State, Kettering University, University of Missouri, and the Lit San Leandro gigabit metropolitan fiber project. The OpenFlow programming primitives of multipath, link aggregation, and goto appear to be sufficient to build multiple paths between two network access points. With redundant paths, communication using our proposed protocol can tolerate faults increased transmission delay due to congestion or to hardware and, thus, reliable communication is achieved. In the event of congestion or failure of a path, the remaining path(s) support(s) communication between the access points while a replacement path is built to reestablish the desired level of fault tolerance. Further investigation and development is needed to establish the precise capability of the OpenFlow Switch Specification version 1.1.0 to support the proposed communication style. It may be that the switch specification would need to be revised. New methods of managing for or responding to congestion, attacks, and other real world challenges will likely need to be developed. Assistance in this investigation and development is welcomed.

Further application information

Additional supporting information, materials and resources

Read about project updates - project blog

Take a look at the existing code - project repository

Will your work be beta-ready by the end of the Development Challenge?


How much effort do you expect this work to take?

This is a challenging project that will require considerable time and effort to achieve a reference level implementation of the protocol. However, we also believe and plan to develop the capability incrementally, with frequent milestones. We have constructed our initial development roadmap and are writing code for our first demonstration. We have three Ph.D. students working at Purdue. There are two faculty and two students working at Kettering at this time, with the expectation that many more will become involved.

Do you need help?

We believe that this team, along with the experimenter support provided by the GENI Project, is sufficient.

If you can help let them know in the comments below.

George Adams

and team members

Dr. George B. Adams III, Team Leader and Director of ManufacturingHUB, Purdue University, Prof. Douglas E. Comer, Distinguished Professor, Department of Computer Science, Purdue University and Ph.D, and graduate students at Purdue to be named. Dr. Adams leads this project and ManufacturingHUB, a solution provider to the National Digital Engineering and Manufacturing Consortium (NDEMC). NDEMC is the first large‐scale public‐private partnership of the United States Government, Original Equipment Manufacturers (OEMs), state and university computing centers, the State of Ohio, and other non‐governmental organizations to provide education, training, and access to computing resources for the Small and Medium‐Size Enterprises (SME’s) manufacturing workforce to develop modeling and simulation skills. With his Ph.D. advisor he conducted research into fault-tolerant networking for parallel computers leading to U.S. patent 4,532,273 that was an inspiration for this project. Dr. Adams is an approved Project Lead of the National Science Foundation-sponsored Global Environment for Network Innovation (GENI) Project and will use this privilege and GENI infrastructure to conduct software defined networking experiments in support of development of the Reliable Communication Protocol. Professor Comer is an internationally recognized expert on computer networking and the TCP/IP protocols. He has been working with TCP/IP and the Internet since the late 1970s. Comer established his reputation as a principal investigator on several early Internet research projects. He served as chairman of the CSNET technical committee, chairman of the DARPA Distributed Systems Architecture Board, and was a member of the Internet Activities Board (the group of researchers who built the Internet). Comer has created courses on computer networks, the Internet, TCP/IP protocols, and operating systems for a variety of audiences, including in-depth courses for engineers and less technical courses for others; he continues to teach at various industries and networking conferences around the world. In addition, Comer consults for private industry on the design of networks and networking equipment. Professor Comer is well-known for his series of ground breaking textbooks on computer networks, the Internet, computer operating systems, and computer architecture. His books have been translated into sixteen languages, and are widely used in both industry and academia. Comer's three-volume series Internetworking With TCP/IP is often cited as an authoritative reference for the Internet protocols. More significantly, Comer's texts have been used by fifteen of the top sixteen Computer Science Departments listed in the U.S. News and World Report ranking. Comer's research is experimental. He and his students design and implement working prototypes of large, complex systems. The performance of the resulting prototypes are then measured. The operating system and protocol software that has resulted from Comer's research has been used by industry in a variety of products. Professor Comer will lead the experimental work of this project. Lit San Leandro: Dr. J, Patrick Kennedy, Principal Investigator and President of OSIsoft; Judi Clark, LitSanLeandro The Lit San Leandro Project Dr. J. Patrick Kennedy, founder of local software company OSIsoft LLC, has taken the lead to modernize the digital infrastructure in City of San Leandro–partly for OSIsoft’s operational needs and partly for the economic development of San Leandro. OSIsoft is the Loop’s first customer. This project will provide the essential elements of a communications infrastructure that, when combined with traditional infrastructure, is critical to businesses that make and sell products to add value and create jobs. This investment will require minimal public investment and will ensure that OSIsoft, one of the largest employers and highest payrolls in San Leandro (and growing), will remain in San Leandro. Additionally, it helps to provide the foundation necessary to attract a new generation of companies to San Leandro, specifically software, technology and green-tech companies that may have a manufacturing component. A Public-Private Partnership Lit San Leandro and San Leandro Dark Fiber LLC comprise the private partnership that work with the City of San Leandro to create the Fiber Loop. San Leandro Dark Fiber owns the fiber optic cable that runs through the City’s underground conduit. Lit San Leandro owns and operates the switch and routing facilities that bring lightning-fast Internet service to our community. Lit San Leandro will serve as a testbed for the project. Kettering University Prof. John Geske, Professor and Head, Department of Computer Science, Kettering University, Flint, Michigan and faculty and undergraduate students to be named. Professor Geske, his faculty to be named, and their undergraduate students to be named will use their soon to be installed NEC GENI rack in support of the experimental component of this project. Students will program the GENI rack, conduct experiments, gather data, and analyze the behavior of RRTP as it is developed. They will interact with Ph.D. students at Purdue and Ohio Supercomputer Center on an occasional basis. Prof. Geske is an approved Project Lead of the National Science Foundation-sponsored Global Environment for Network Innovation (GENI) Project and will use this privilege and GENI infrastructure to conduct software defined networking experiments in support of development of the Reliable Communication Protocol. In particular, he will coordinate protocol development that uses the GENI equipment installed at Kettering University. Here follows Professor Geske’s remarks on participation in this project: “Flint’s Charles Stewart Mott Foundation provided US Ignite with a planning grant to help kick-start US Ignite’s non-profit partnership and has helped engage other national and local partners to bring the technology benefits to Flint. Mott believes US Ignite has the potential to help people improve their lives and their communities and enhance the work in the areas of clean energy, advanced manufacturing, and education. The Mott Foundation has designated Kettering University as the hub for US Ignite collaborative efforts for the Flint region, and the networking infrastructure necessary to support these initiatives is currently being developed. As such, Kettering University is committed to the creation of collaborative efforts that support the articulated vision of the Mott Foundation. The proposed collaboration is completely aligned with this vision, and it is a significant step forward in Kettering University achieving the stated goals of US Ignite. Kettering University (formerly General Motors Institute for Engineering and Management), is the pre-eminent undergraduate science and engineering institution incorporating experiential learning into engineering and science education. All students at Kettering University are required to have a minimum of five experiential terms. These experiences may be from main-line manufacturers, (e.g., Ford, GM, Bosch, Siemens), healthcare (e.g., Medtronic, GE Medical, Epic, Ascension Health), service industries (Auto-Owners, Plex Systems, E-Prize), and government agencies (e.g., NSA, CIA, FBI, Argonne, NIST). As a result, the typical Kettering University student has an enriched academic experience that couples foundational principles with “real-world” implementations and applications. “The Kettering University Computer Science Department (an ABET-accredited computer science program) is committed to providing faculty and undergraduate students for this collaboration. Faculty with a broad range of academic and industry experience in networking, web-development, software engineering, and real-time systems will provide their expertise and guidance, and will supervise undergraduate student researchers pursuing their academic theses. The Senior Thesis is a unique aspect of a Kettering University education; the thesis project is an academic requirement that is completed during the student’s senior year. It is the culmination of their experiential experiences. The projects that develop from this collaboration will enhance their practical experiences, and provide them with research and development experiences that can immediately be transferred into their workplaces.”

comments powered by Disqus