Timely Status Update in Massive IoT: Regular Delivery and AoI Optimization

Timely Status Update in Massive IoT

The freshness, or timeliness, of the status update information maintained at interested nodes is critical in many real-world IoT (Internet of Things) applications. To characterize it, a new performance metric called age-of-information (AoI) has been proposed and its optimization with low signaling overhead is of particular interest. Towards this end, the following questions are posed: How to optimally schedule the packets to ensure regular packets delivery? What is the optimal terminal scheduling policy to minimize AoI? Can we solve for the closed-form optimal performance? Is there a decentralized policy, i.e., terminals transmit based on its own state information without signaling exchange, that can achieve universally optimum?

Optimal Scheduling for Regular Delivery^[1,2]

To guarantee the regularity of the inter-deliver time, the value iteration method can be applied to find the optimal policy. However, the bane of this and many similar problems is the resulting complexity (curse of dimensionality). In an attempt to make fundamental progress, we take a high-reliability asymptotic approach, in which the channel failure probabilities for different clients are of the same order and asymptotically approach to zero. Under this circumstance, we show that the asymptotically optimal policy is a “modified least time-to-go” policy.

Optimal Scheduling for AoI Minimization^[3,4]

If the status update packets are generated stochastically, they should be scheduled to minimize their average AoI. Toward this end, we have proposed a Whittle’s index policy and derived the index in closed-form and established the indexability thereof. Then, we further propose a round-robin policy with one-packet (schedule the latest packet only and drop the others) buffers (RR-ONE) and prove that RR-ONE is asymptotically optimal among all policies in the massive IoT regime. The steady-state stationary distribution of AoI under RR-ONE is also derived.
However, the Whittle’s index policy is in nature a centralized scheduling policy, requiring global information from every clients. In this regard, we further propose an index-prioritized random access (IPRA) policy, which achieves the universally near-optimal AoI performance but with decentralized protocol structure. Specifically, a client whose index is higher than a threshold will be scheduled to send with a probability p; otherwise it waits.

[1] X. Guo, R. Singh, P. R. Kumar and Z. Niu, "A High Reliability Asymptotic Approach for Packet Inter-Delivery Time Optimization in Cyber-Physical Systems", ACM MobiHoc’15, 2015.
[2] X. Guo, Z. Niu, S. Zhou, P. R. Kumar, "A Risk-Sensitive and High Reliability Asymptotic Approach for Packet Inter-Delivery Time Optimization in Cyber-Physical Systems", IEEE/ACM Trans. Networking, 2018 (under revision)
[3] Z. Jiang, B. Krishnamachari, S. Zhou, Z. Niu, “Can Decentralized Status Update Achieve Universally Near-Optimal Age-of-Information in Wireless Multiaccess Channels?”, submitted to Intl. Teletraffic Congress (ITC), 2018
[4] Z. Jiang, B. Krishnamachari, X. Zheng, S. Zhou, Z. Niu, “Decentralized Status Update for Age-of-Information Optimization in Wireless Multiaccess Channels”, to be presented at IEEE Intl. Symp. Info. Theory, 2018