university of illinois at urbana champaign department of computer science 1 university of illinois...

University of Illinois at Urbana Champaign

Department of Computer Science

1



1

A Brief History

of

Networked Real Time Embedded System

Laboratory



2

Schedulability Analysis The Networked Real-Time Embedded System Laboratory

(NRTESL) at UIUC has a long and distinguished record on fundamental research with great impacts on the engineering of real-time embedded systems.

The laboratory was founded by David Liu and Jane Liu in the 70’s. Professor David C. L. Liu analyzed the basic properties of the two mostly widely used scheduling algorithms, the rate monotonic and deadline scheduling algorithms for independent periodic tasks. The Rate Monotonic Scheduling algorithm was used in the NASA Apollo program http://history.nasa.gov/apollo.html

His seminal work pioneered schedulability analysis - It is a foundation for many of the subsequent research in real-time computing. Among many other awards, Prof. Liu received the 1st Outstanding Technical Achievement Award from the IEEE Real-Time System Committee in 1999 for his seminal work on real-time computing.



3

Imprecise Computation During late 80’s and early 90’s, Professors Jane Liu and Kwei Jay Lin

pioneered the research on the imprecise computation technique which allows for flexible trade-offs between timeliness and precision and is a effective means to handle overload conditions and provide graceful degradation in real-time environments.

They and their students developed algorithms for scheduling imprecise tasks to trade quality for time and fault tolerance strategies based on this technique.

To enable its use in application domains as diverse as databases, real-time control and data communication, they developed a relational query processor that returns partial answers with increasing accuracy as more data is retrieved and processed, determined the quality/time tradeoffs of direct digital controllers, and evaluated the effectiveness of this technique for multimedia traffic congestion control.



4

Open Real Time Computing Architecture

Professor Jane Liu's other work include algorithms for end-to-end scheduling in multiprocessor and distributed systems and rigorous and tractable methods to bound the effect of timing anomalies exhibited by tasks in these systems.

In the early 90's, she led the effort in the development of PERTS (Prototyping Environment for Real-Time Systems). This system of schedulers and tools enables its users to experiment with alternative real-time resource management strategies, to synthesize and tune the run-time system, and to validate the timing properties of the resultant system. PERTS is the earliest of such tools and is still the most widely marketed and used one.

She and her students developed the Open Real-Time Computing architecture and its underlying principles. This architecture allows independently developed real-time and non-real-time applications to effectively share computing and communication resources without interfering each other.



5

Generalized Rate Monotonic Scheduling

Professor Lui Sha joined Laboratory in 1998. Building upon the pioneering work by David C. L. Liu, he worked with John Lehoczky and Ragunathan Rajkumar at CMU and with John Goodenough and Mark Klein at SEI developed and transitioned into practice a comprehensive real-time scheduling theory known as Generalized Rate Monotonic Scheduling (GRMS). Professor Sha was elected ACM Fellow in 2006 “for contributions to real time systems” and elected IEEE Fellow in 1998 for “for technical leadership and research contributions which enabled the transformation of real-time computing practice from an ad hoc process to an engineering process based on analytic methods.”

GRMS is now the foundation for the real-time computing in IEEE Futurebus+, IEEE POSIX RT extension, Ada 95, real-time CORBA, real-time Java, real-time UML. These real-time computing infrastructure have become the basic building blocks of modern real-time systems. GRMS was cited in the Selected Accomplishment Section (p. 193) of the National Research Council's report, A Broader Agenda for Computer Science and Engineering, 1992.



6

“Through the development of Rate Monotonic Scheduling [theory], we now have a system that will allow [Space Station] Freedom's computers to budget their time, to choose between a variety of tasks, and decide not only which one to do first but how much time to spend in the process”

Deputy Administrator of NASA, Aaron Cohen, “Charting The Future: Challenges and Promises Ahead of Space Exploration”, pp. 3, October 1992.

International Space Station



7

“The navigation payload software for the next block of Global Positioning System upgrade recently completed testing. ... This design would have been difficult or impossible prior to the development of rate monotonic theory”, L. Doyle, and J. Elzey, “Successful Use of Rate Monotonic Theory on A Formidable Real-Time System”, in the Proceedings of 11th IEEE Workshop on Real-Time Operating Systems and Software, pp. 74, 1994.

NOMAD on its way through the

Antarctic.



8

“The Mars Pathfinder mission was widely proclaimed as "flawless" in the early days after its July 4th, 1997 landing on the Martian surface. … But a few days into the mission, not long after Pathfinder started gathering meteorological data, the spacecraft began experiencing total system resets… Once diagnosed, it was clear to the JPL engineers that using priority inheritance would prevent the resets they were seeing. …No more system resets occurred.

David also said that some of the real heroes of the situation were some people… who first identified the priority inversion problem and proposed the solution …They were Lui Sha, John Lehoczky, and Raj Rajkumar. …

When was the last time you saw a room of people cheer a group of computer science theorists for their significant practical contribution to advancing human knowledge? :-) It was quite a moment.”

http://catless.ncl. ac.uk/Risks/ 19.49.html



9

Building Robust Systems Professor Lui Sha led the development of Simplex architecture for

building robust real time control systems. The architecture principles include the use of GRMS for timing correctness, real time publication and subscription to decouple system components, the use of simple and analytically redundant components to ensure critical functions, explicit and machine checkable environmental assumptions, and the enforcement of well-formed dependency.

“Lui’s contribution is immeasurable. His superior knowledge in detailed architecture design attributes, fault tolerant strategies, and software development practices were a tremendous aid in calibrating the F35 Mission Systems Design Philosophy. His support has been nothing short of flawless in accessing complex advanced avionics.” John L. Hudson, Major General, USAF, JSF Program Executive Officer, April 5, 2004

“Prof Sha’s assistance has been invaluable in raising software stability by an order of magnitude on the F/A-22 lead jets. Prof Sha has played a central role in assisting the F/A-22 program to improve its avionics stability.” Dr. Andre vanTilborg, Director of Information Technology, Office of the Secretary of Defense, 9/24/2003



10

“The bottom line is that now we are basing the mathematical Analysis on work done by Dr. Sha.” Greg Marin, Rockwell Collin March 2001. Report to FAA on the analysis of timing behavior of Real Time Virtual Processors for DO 178B level A certification.



11

Major Wireless Network Accomplishments

• We are the first to systematically study performance limits of wireless sensor networks with respect to network coverage, connectivity, lifetime, critical power.

• We have devised the first localized and distributed topology control algorithm and rigorously its connectivity preservation property.

• We have designed and implemented (on top of Atheros chipsets) an IEEE 802.11e-based protocol that provides proportional deterministic QoS.

• Jointly with Champaign Urbana Wireless Mesh Networks, we have designed and deployed a city-wide wireless mesh network that serves both as a research testbed and a production network.

• On CuWIN, we have designed and implemented transparent device drivers that provide a rich set of APIs for cross-layer optimization.

Wireless Network Simulator (J-Sim)

• Network capacity• Connectivity• Fairness and QoS

Physical Wireless Testbed

Wireless Network Emulator

•Frequent disconnections due to mobility and wireless erros• Time-varying channels due to environmental factors• Intra-/Inter-flow interference• Energy constraints•



12

Personal Assistant System for HealthCare We have laid a software infrastructure, PAS, that exploits inexpensive, "off the shelf“

technologies to assist elderly people to maintain the capability of independent living through time-based reminders of daily activities, non-intrusive monitoring of physiological functions and mobility profiles, and real-time communications with remote care providers and clinicians.

• A network of small, low-power devices that integrates radio-frequency identification (RFID) readers/tags, sensors, and bluetooth-enabled medical devices.

• Capable of tracking in real-time (in addition to locating) humans and objects with the combination of ultrasonic and RFID technologies.

• Leveraging cell phones as both the wireless modem and the local intelligence for data aggregation and acquisition.

We are currently working with Geriatricians at Washington University in Saint Louis in evaluating operations of a PAS prototype in Nazareth Living Center for Assisted Living with patients with diverse medical needs.

Monitoring Service

Clinician Service

Gateway

Home Environment

MedicalDevice



13

JavaSim: Large-Scale Network Simulation/Emulation

Composability• Allow study of different networking problems in

a unified, coherent framework. Extensibility

• Allow code reuse and insertion of new codes so as to accommodate new architectures and services.

Level of abstraction• Allow simulation conducted at different levels of

abstraction and at different granularities. Diagnosis and monitoring

• Allow diagnosis and monitoring to be conducted at the component level in a time efficient manner.

Network emulation.• Allow real-life applications to run on J-Sim.

High fidelity in terms of• Physical layer characteristics• MAC through transport layer protocols• Mobility models• Network connectivity models• Real vs. virtual environments

virtu

al

real



14

Efficient overload management in real-time computing

Professor Marco Caccamo joined the Laboratory in 2002. He developed a comprehensive CPU scheduling framework for managing unpredictable overruns in a real-time environment, where tasks may have different criticality, flexible timing constraints, shared resources and variable execution times. The proposed methods (capacity and bandwidth sharing) provide an efficient resource sharing mechanism that reclaims unused bandwidth, provides temporal isolation, and enhances task responsiveness. This theory has been generalized to the case of multi-processor real-time platforms and real-time wireless LANs. These scheduling methodologies greatly enhance efficiency of EDF scheduling on single-CPU and multi-processor platforms.



15

Implicit EDF for robust RT wireless communication

Professor Marco Caccamo's other work includes algorithms for predictable real-time wireless communication. Existing wireless networks lack adequate real-time support for multi-media traffic and emergency response. Temporal predictability is a serious concern especially in large scale deployments due to unbounded priority inversion problem and packet collisions.

Without making idealistic assumptions on operating environment and channel quality, Caccamo and his research group devised a suite of predictable real-time wireless protocols based on the notion of “implicit contention”. For instance, Robust Implicit EDF (RI-EDF) is a high performance protocol for distributed sensing and control: formations of collaborative robots/UAVs can take advantage of its robustness and temporal isolation. For sensor networks, Caccamo and his students have introduced a scheme that allows the coexistence of non-real-time traffic (CSMA/CA of IEEE802.15.4) and prioritized real-time traffic for emergency response.



16

Looking Ahead

In 2006, the laboratory was renamed as Networked Real Time and Embedded System Laboratory, dedicated to the mission of create the science and technology foundation and accompanying curriculum to foster the convergence of • computing, • communication • and the intelligent control of the physical environment.

university of illinois at urbana champaign department of computer science 1 university of illinois...

Documents