Abstract
The surging demand for fresh information from various Internet of Things (IoT) applications requires oceans of data sampled from the physical environment to be transmitted and processed timely, which would lead to huge energy consumption. We investigate the sleep-wake strategies of servers in communication systems to reduce energy consumption while guaranteeing timely delivery of fresh information to users. Specifically, we investigate a multi-source single-server queueing system and propose a novel sleep-wake strategy called the Conditional Sleep (CS) scheme. Our analysis reveals that the CS scheme outperforms the widely-used Hysteresis Time (HT) and Bernoulli Sleep (BS) schemes in terms of Age of Information (AoI), while retaining the same energy consumption rate and Peak Age of Information (PAoI). We find that increasing the sleep period length leads to a reduction in energy consumption and an increase in PAoI, but it does not always increase AoI. Moreover, we show that using PAoI as the information freshness metric in designing sleep-wake strategies would make the server sleep infinitely long due to the PAoI being determined by first-order statistics. We further numerically show that having the bufferless system can achieve a better PAoI-energy tradeoff than the infinite buffer system when having a large sampling rate.
Introduction
In the Internet of Things (IoT), information "freshness" is critical. For applications like autonomous driving or industrial monitoring, stale data can lead to system failures. The primary metric for freshness is the Age of Information (AoI), defined as the time elapsed since the last received update was generated. However, maintaining low AoI requires frequent sampling and processing, which is energy-intensive. This paper explores the tradeoff between energy consumption and information freshness in server systems that utilize sleep-wake strategies to save power.
System Model and Proposed Strategy
The authors model a multi-source, single-server, bufferless queueing system. A bufferless approach is often superior for freshness because it prioritizes the newest arrivals over stale, backlogged data, preventing a "backlog" of old information.
The paper introduces a novel Conditional Sleep (CS) scheme and compares it against two traditional methods:
- Hysteresis Time (HT): The server waits for a fixed duration after processing; if no new packet arrives, it enters sleep mode.
- Bernoulli Sleep (BS): The server decides to sleep with a certain probability after each task.
- Conditional Sleep (CS): The server sleeps only if the processing time of the previous packet was below a certain threshold.
The core idea of CS is to avoid sleeping after long processing tasks, which already leave the system with a high AoI. By sleeping only after "fast" tasks, the CS scheme reduces the variance in update intervals, which is the secret to lowering average AoI.
Key Insights and Analysis
The analysis employs renewal-type methods to derive closed-form expressions for AoI, Peak AoI (PAoI), and energy consumption.
The Role of Variance
A critical finding is that while energy consumption and PAoI depend on first-order statistics (average sleep cycle length), the Average AoI depends on second-order statistics (variance). The CS scheme effectively reduces this variance, allowing it to achieve a lower AoI than HT or BS for the same energy budget.
Monotonicity and Metrics
The behavior of the system under the CS scheme varies with parameters like the sleep period length ($N$) and the threshold ($b_i$).
As shown in the table above, increasing the sleep period ($N$) decreases energy consumption but has a non-monotone effect on Average AoI. Interestingly, optimizing for PAoI alone can lead to an "infinite sleep" trap, where the server sleeps indefinitely because PAoI does not sufficiently penalize the large age gaps created during sleep as effectively as Average AoI does.
Key Results
- Superior Freshness: In multi-source environments, the CS scheme can reduce AoI by up to 30% compared to the HT scheme at the same energy level.
- Significant Energy Savings: Sleep-wake strategies can cut energy consumption by 35% to 65% compared to an always-on server.
- Bufferless Advantage: At high sampling rates, bufferless systems provide a better age-energy tradeoff than infinite buffer systems.
Conclusion
The Conditional Sleep strategy offers a robust way to balance energy efficiency with information freshness in IoT servers. By leveraging second-order statistics, it outperforms traditional idling schemes. Future work may extend these findings to multi-server systems and explore cooperative sampling behaviors between sources and servers.
Summary updated