Ultra-low latency and High Reliability for Wireless IoT

The goal of this project is to develop a new framework for supporting ultra-low latency and high reliability network services for emerging Internet-of-Things applications.  These emerging applications often impose stringent requirements on both latency and reliability that cannot be met with existing techniques.  The main idea behind this project is that latency and reliability are inherently coupled and must be addressed not just at the physical layer but also at higher network layers.  Our comprehensive research agenda addresses latency and reliability at various time scales, ranging from packet time scale and MAC layer scheduling to network deployment and the economics of their services.  Our proposed research includes the following interrelated tasks:

  • Packet time scale optimization: Issues here include how messages are divided into packets, the length of time-slot used to send packets, and various physical layer parameters.
  • MAC time-scale optimization: The focus here is on scheduling traffic across different flows so as to meet the latency and reliability requirements of IoT tasks. 
  • Network level time-scale:  The focus here is on end-to-end reliability over a multi-hop network, by exploiting network diversity to ensure an adequate level of reliability.
  • Network planning time-scale:  Issues here include how networks are deployed and how sharing agreements are formed among network operators to enhance network diversity and reliability.

This project is a collaboration between Northwestern University, MIT, and the University of Oulu, Finland.

Major Activites:

1.  PI Randall Berry gave a tutorial at the 2018 WiFiUS summer school also gave a webinar aspart of the IEEE 5G initiative – both of which included some of the results produced inthis project.

2. The team held a face-to-face meeting in Boston, in June of 2018.

3.  PI Modiano and Glisic met in person in November, 2019.

Significant Results:

1. Bundling Licensed and Unlicensed Spectrum:

X. Wang and R. Berry, “The impact of bundling licensed and unlicensed wireless services,” IEEE INFOCOM 2018.

Unlicensed spectrum has proven to be a key enabler for new wireless services including IoT. However, due to its open nature, services deployed in unlicensed spectrum may suffer from unreliability and latency. One way to mitigate this is for a service provider to combine unlicensed access with licensed access and then sell this “bundled service” to end users. This bundling can also impact the competition of service providers for customers. In prior work we have studied service provider competition without bundling using game theoretic models for competition with congestible resource developed in the operations and economics literature. We have extend these models to allow for bundling and have been able to show that the use of bundling results in significantly different market equilibria and can improve overall welfare.

X. Wang and R. Berry, “Price Competition with LTEU and WiFi,” IEEE INFOCOM 2019.

LTEU is a variation of LTE the operates in unlicensed spectrum and can be used to facilitate the type of bundling discussed above. However, at the MAC layer thisn technology is based on a different approach than that used in WiFi,the dominant technology deployed in unlicensed spectrum. There has been work considering the coexistence of these two technologies from a technical perspective. Our work has instead focused on the economic dimension of this interactions. Namely we have extended our framework for studying competition between service providers to account for providers using different technologies and characterized the impact of these technologies on this competition.

2. Contracts, investment and shared spectrum

Y. Zhu, and R. Berry, “Contracts as investment barriers in unlicensed spectrum,” IEEE INFOCOM 2018.

Another way to reduce the potential for high latency and low reliability in unlicensed spectrum is through the use of longterm contracts, which in effect limit the ability of new service providers to enter a market and further congest the spectrum. However such contracts may also lower a firm’s incentives to invest in the development and  deployment of new technology. We have studied such issues by using multistage game theoretic models in which an incumbent firm offers a longterm contract to users who are uncertain about future potential QoS offered by an entrant.

Y. Zhu, H,. Yu and R. Berry, “CrossNetwork Prioritized Sharing: An Added Value MVNO’s Perspective” IEEE INFOCOM 2019

Another form of contract that arises in wireless markets is between a service provider and a Mobile Virtual Network Operator (MVNO). Traditionally MVNO’s operated bycontracting with a single service provider and using their network to offer services.  However, recently there are examples of MVNOs that contract with multiple service providers and use more thanone network to offer service. We are studying such scenarios to gain insight into the potential gains and losses from such crossnetwork sharing.

B. Lorenzo, Savo Glisic, et.al. Data and Spectrum Trading Policies in a Trusted Cognitive Dynamic Network Architecture IEEE/ACM Transactions on Networking , 2018.

Future wireless networks will progressively displace service provisioning towards the edge to accommodate increasing growth in traffic. This paradigm shift calls for smart policies to efficiently share network resources and ensure service delivery. We consider a cognitive dynamic network architecture (CDNA) where primary users (PUs) arerewarded for sharing their connectivities and acting as access points for secondary users (SUs). CDNA creates opportunities for capacity increase by networkwide harvesting of unused data plans and spectrum from different operators. Different policies for data and spectrum trading are presented based on centralized, hybrid, and distributed schemes involving primary operator (PO), secondary operator (SO), and their respective end users.

 

3. Markets for spectrum monitoring

A. Ghosh, R. Berry and V. Agrawal, “Spectrum Measurement Markets for Tiered Spectrum Access,“ IEEE Transactions on Cognitive Communications and Networking,2018.

A promising way increasing the amount of spectrum available new low latency/high reliability applications is through sharing spectrum on a secondary basis with federalincumbents. A key approach to facilitate this is through measuring spectrum usage so as to avoid interfering with incumbent users. This is the role of the so called ESC operators in the newly established CBRS band (3.5 GHZ). We consider a scenario in which multiple ESC’s compete to provide spectrum access data to service providers. We study how differences in the quality and price of these measurements impact the resulting market dynamics among the service providers. Our work has revealed some unexpected results such as having ESC operators with different qualities may be beneficial. However these conclusions also depend on how service providers share the underlying spectrum.

A. Shams, S. Glisic, et. al., User Centric Distributed Spectrum Sharing in Dynamic Network Architectures , IEEE/ACM Transactions on Networking , 2018.

In this work we develop and analyze a new usercentric networking model for ubiquitous spectrum sharing where every user can share and use the spectrum underuncertainty of their traffic models. In this concept, users when connected to the Internet (wired/wireless) can dynamically serve as access points for other users in their vicinity. The model is designed as an operator supervised doubleStackelberg game with network operators, access points, and users as main players. We study network reliability and latency of the system under uncertainty of users’ traffic patterns.

4. Low latency Networking

I. Kadota, A. Sinha, E. Modiano, "Optimizing Age of Information in Wireless Networks with Throughput Constraints," IEEE Infocom 2018. (Best paper award winner)

Age of Information (AoI) is a performance metric that captures the freshness of the information from the perspective of the destination. The AoI measures the time that elapsed since the generation of the packet that was most recently delivered to the destination. In this paper, we consider a singlehopwireless network with a number of nodes transmitting timesensitive information to a Base Station and address the problem of minimizing the Expected Weighted Sum AoI of the network while simultaneously satisfying timelythroughput (reliability) constraints from the nodes.

R. Talak, S. Karaman, E. Modiano, "Optimizing Information Freshness in Wireless Networks under General Interference Constraints," ACM MobiHoc 2018. (Best paper award winner).  

We consider the problem of minimizing average and peak AoI in wireless networks under general interference constraints. When fresh information is always available for transmission, we show that a stationary scheduling policy is peak age optimal. We also prove that this policy achieves average age that is within a factor of two of the optimal average age.

V. Tripathi, R. Talak, E. Modiano, "Age Optimal Information Gathering and Dissemination on Graphs," IEEE Infocom, 2019.

We consider the problem of timely exchange of updates between a central station and a set of ground terminal via a mobile agent that traverses across the ground terminals along a mobility graph and design the trajectory of the mobile agent to minimize peak and average age of information.

S. Shafigh, S. Glisic and B. Lorenzo, "Dynamic Network Slicing for Flexible Radio Access in Tactile Internet," GLOBECOM 2017.

We propose a flexible cloudbased radio access network (FRAN) for TI, where traffic of user equipments (UEs) can be temporary offloaded from the operatorprovided networks to user provided networks.

 

Rajat Talak, Sertac Karaman, Eytan Modiano, “When a Heavy Tailed Service Minimizes Age of Information,” IEEE International Symposium on Information Theory, Paris, France, July, 2019.

We show that AoI and packet delay differ in a fundamental way in certain systems, i.e. minimizing one can imply maximizing the other. We consider two queueing systems, namely a single server last come first serve queue with preemptive service (LCFSp) and G/G/inf queue, and show that a heavy tailed service distribution, that results in the worst case packet delay or variance in packet delay, respectively, minimizes AoI. For the specific case of M/G/1 LCFSp and G/G/inf queue, we also prove that deterministic service, that minimizes packet delay and variance in packet delay, respectively, results in the worst case AoI.

 

Rajat Talak, Eytan Modiano, “Age-Delay Tradeoffs in Single Server Systems,” IEEE International Symposium on Information Theory, Paris, France, July, 2019.

We prove that there is a natural tradeoff between the AoI and packet delay. We consider a single server system, in which at most one update packet can be serviced at a time. The systemdesigner controls the order in which the packets get serviced and the service time distribution, with a given service rate. We analyze two tradeoff problems that minimize packet delay and the variance in packet delay, respectively, subject to anaverage age constraint. We prove a strong age-delay and agedelay variance tradeoff, wherein, as the average age approaches its minimum, the delay and its variance approach infinity.

 

Rajat Talak, Sertac Karaman, Eytan Modiano, “A Theory of Uncertainty Variables for State Estimation and Inference,” Allerton Conference on Communication, Control, and Computing, September 2019.

We develop a new framework of uncertainty variables to model uncertainty in the real world. A simple uncertainty variable is characterized by an uncertainty set, in which its realization is bound to lie, while the conditional uncertainty is characterized by a set map, from a given realization to a set of possible realizations of another variable. We prove Bayes’ law and the law of total probability equivalents for uncertainty variables. We then develop a graphical model, namely Bayesian uncertainty network, a Bayesian network equivalent defined over a collection of uncertainty variables, and show that all the natural conditional independence properties, expected out of a Bayesian network, hold for the Bayesian uncertainty network. We also define the notion of point estimate, and show its relation with the maximum a posteriori estimate.

 

Vishrant Tripathi, Eytan Modiano, “A Whittle Index Approach to Minimizing Functions of Age of Information,” Allerton Conference on Communication, Control, and Computing, September 2019.

We consider a setting where multiple active sources send real-time updates over a single-hop wireless broadcast network to a monitoring station. Our goal is to design a scheduling policy that minimizes the time-average of general non-decreasing cost functions of Age of Information. We use a Whittle index based approach to find low complexity scheduling policies that have good performance, for reliable as well as unreliable channels. We also derive structural properties of an optimal policy, that suggest that the performance of the Whittle index policy may be close to optimal in general. These results might also be of independent interest in the study of restless multi-armed bandit problems with similar underlying structure. 

 

 

Broader Impacts 

Improving the robustness of spectrum and infrastructure sharing mecahnisms will enable these techniques to be commercially viable and improve wireless access worldwide.  As part of WiFiUS, the project will strengthen ties between networking researchers in the US and Finland. Other broader impacts will include: (1) dissemination of the results through journal and conference publications to advance the field; (2) incorporation of new theory and algorithms developed through this work into advanced graduate classes at MIT, NWU and OU (3) development of experimental projects for undergraduate students (4) technology transfer to industry and government laboratories; (5) training of graduate and undergraduate students.

This project is training a number of graduate students in the information technology field, which is of critical importance. In particular, this project has trained a number of graduate students in the area of wireless networks and resource allocation. At MIT, two graduate students completed their Master's degrees working on this project, at Northwestern University one (female) graduate student is pursuing a PhD under this project and at the University of Oulu one student completed a PhD under this project.

 

People

PIs:  

Randall Berry, Northwestern University,

Eytan Modiano, MIT

Savo Gilsic, University of Oulu

 

Students:

Ertem Tas

Xinzhe Fu

Vishrant Tripathi

Rajat Talak

 

Publications:

  1. Thomas Stahlbuhk, Brooke Shrader, Eytan Modiano, “Throughput Maximization in Uncooperative Spectrum Sharing Networks,”  IEEE International Symposium on Information Theory, Barcelona, Spain, July 2016.
  2.  Qingkai Liang, H.W. Lee, Eytan Modiano, “Robust Design of Spectrum-Sharing Networks,” IEEE Wiopt, Tempe, Arizona, May, 2016.
  3. S. Shafigh, S. Glisic and B. Lorenzo, "Dynamic Network Slicing for Flexible Radio Access in Tactile Internet," GLOBECOM 2017.
  4. X. Wang and R. Berry, “The impact of bundling licensed and unlicensed wireless services,” IEEE INFOCOM 2018.
  5. Y. Zhu, and R. Berry, “Contracts as investment barriers in unlicensed spectrum,” IEEE INFOCOM 2018.
  6. R. Talak, S. Karaman, E. Modiano, "Optimizing Information Freshness in Wireless Networks under General Interference Constraints," ACM MobiHoc 2018.
  7. A. Ghosh, R. Berry and V. Agrawal, “Spectrum Measurement Markets for Tiered Spectrum Access,“ to appear IEEE ICC, 2018.
  8.  Igor Kadota, Abhishek Sinha, Eytan Modiano, "Optimizing Age of Information in Wireless Networks with Throughput Constraints," IEEE Infocom, April 2018.
  9. Vishrant Tripathi, Rajat Talak, and Eytan Modiano, Age Optimal Information Gathering and Dissemination on Graphs, IEEE Infocom, 2019.
  10. Rajat Talak, Sertac Karaman, Eytan Modiano, “When a Heavy Tailed Service Minimizes Age of Information,” IEEE International Symposium on Information Theory, Paris, France, July, 2019.
  11. Rajat Talak, Eytan Modiano, “Age-Delay Tradeoffs in Single Server Systems,” IEEE International Symposium on Information Theory, Paris, France, July, 2019.
  12. Rajat Talak, Sertac Karaman, Eytan Modiano, “A Theory of Uncertainty Variables for State Estimation and Inference,” Allerton Conference on Communication, Control, and Computing, September 2019.
  13. Vishrant Tripathi, Eytan Modiano, “A Whittle Index Approach to Minimizing Functions of Age of Information,” Allerton Conference on Communication, Control, and Computing, September 2019.