Finite-blocklength coding (FBC) is a promising technology to achieve ultra-reliable and low-latency communications (URLLC) in emerging applications, e.g., autonomous driving and extended reality. The challenge lies in scheduling random arriving traffic in URLLC due to the blocklength constraint imposed by low latency. This paper designs a cross-layer mechanism, called the joint scheduling and FBC policy, to meet URLLC requirements for bursty traffic under AWGN and block fading channels. First, we model a single-user transmission system as a tandem queue model. In this model, a packet queue buffers randomly arriving packets. After these packets are encoded using FBC, the resulting encoded symbols are buffered in a symbol queue. To analyze the latency and reliability performance, we represent the system as a Markov chain and conduct steady-state and transition analyses. After that, we construct a non-convex problem to minimize delay subject to reliability and power constraints. Using a variable combination method, we convert this problem into a linear-fractional programming (LFP) problem. Notice that extremely high reliable requirement significantly increases the computational complexity of standard LFP method.We approximate the objective function and packet drop ratio constraint, and obtain a lower bound of the minimum average delay effectively with a marginal performance loss.

Accurate classification of encrypted traffic is vital for effective network management and Quality of Service (QoS). Protocol identification and application recognition are central to this process. Existing solutions generally use separate models for protocol identification and application recognition. However, running these models in parallel in online environments significantly increases system complexity, resource consumption, and maintenance costs. Additionally, these models often suffer from poor traffic representation by discarding crucial byte information and retaining irrelevant biases. This paper introduces an efficient multi-task learning framework, ELEMTC, designed to simultaneously perform protocol identification and application recognition, while offering a tailored traffic representation scheme for network management. To improve efficiency, we propose a novel sample pre-training task, Detection Token Replacement, specifically tailored for the networking domain. Experimental results on the ISCX VPN-non VPN dataset show that ELEMTC achieves Fl scores of 98.37% and 99.46% for application recognition and protocol identification, respectively. These results demonstrate that ELEMTC significantly outperforms current state-of-the-art methods. Furthermore, ELEMTC demonstrates outstanding efficiency by improving inference speed 20 times while maintaining relatively low memory usage.

Achieving high reliability and low latency is a critical challenge for a wide range of applications that demand strict performance guarantees, such as real-time systems, industrial automation, and autonomous vehicles. Our primary focus is on ultra-reliable low-latency communication (URLLC), which aims to ensure a real-time requirement. We propose a solution for hard delay-constrained communication, which meets strict latency requirements by incorporating variable-length coding in short-packet transmission systems. Our approach utilizes a cross-layer design using truncated channel inversion transmission across parallel channels. This system can be characterized as a two-dimensional Markov chain, which consists of both the packet queue buffering bits to be encoded and symbol queue buffering coded symbols to be transmitted. By leveraging embedded Markov chains, we formulate an optimization problem to minimize the average power consumption while converting the problem into a one-dimensional Markov chain. We present a heuristic algorithm to obtain hard delay-constrained policies and utilize gradient descent policy to refine the policies and explore the trade-offs between hard delay constraints and power consumption.

Ultra-Reliable and Low-Latency Communication (URLLC) has attracted considerable attention because it has great potential in industrial automation and autonomous driving. As one of the promising techniques to solve the low-latency problem, finite blocklength coding has been a critical topic. How to reduce the delay with the finite blocklength coding in the URLLC system is a challenging issue. In this paper, a Lyapunov drift-based scheduling scheme is presented for short-packet transmission systems. Specifically, a scheduling scheme is designed to determine the transmission rate and maximize the system throughput with Lyapunov optimization. The average power constraint can be tackled with the virtual power queue method. So that the short-packet transmission problem can be formulated as a univariate optimization problem, which is non-convex. A sub-optimal solution to this optimization problem can be obtained by our presented algorithm, which ignores the higher order terms of the Taylor series. Through simulations, the performance of our proposed algorithm is shown to be very close to that of the optimal policy obtained by the exhaustive search algorithm.

Finite blocklength coding has attracted considerable recent attention because it holds the promise of ultra-reliable and low-latency communications (URLLC) in smart grids, autonomous driving, tele-surgery, and industrial internet of things (IIoT). However, as the instantaneous blocklength is constrained by the number of backlogged bits, short packet transmission with random packet arrival becomes a challenging issue. In this paper, we present a cross-layer mechanism referred to as the buffer-aware variable-blocklength coding to minimize the average delay of bursty traffics. To optimize and analyze the buffer-aware short packet transmission, we formulate a tandem queue model consisting of both the packet queue buffering bits to be encoded and the symbol queue buffering coded symbols to be sent. By deriving the transition probability matrix and the steady state probability of the two-dimensional Markov chain characterizing the tandem queue, we obtain the average latency as a function of the arrival rate and transmission power. Simulation results verify our theoretical analysis and demonstrate the potential of the buffer-aware variable-blocklength coding scheme.

Non-orthogonal multiple access (NOMA) has received considerable attention as a promising candidate for future mobile networks. How to reduce traffic delay through cross-layer scheduling in NOMA systems is a challenging issue. In this paper, a cross-layer approach for NOMA systems is designed to fulfill the delay requirements. In particular, scheduling decisions should adapt to the system dynamics. With the distribution information of the system dynamics, the constrained Markov decision process (CMDP) is employed to characterize the scheduling decision. The optimal scheduling policy can be obtained by converting the scheduling problem to linear programming. A computational approach based on the Kronecker product is conceived to formulate the linear programming when the dimension of the CMDP is high. Then the optimal delay-power tradeoff can be achieved. Moreover, the optimal decoding order is demonstrated that it can be obtained directly based on the channel states. The threshold-based structure of the optimal scheduling decisions is revealed. Without the distribution information, the Lyapunov approach is exploited to propose an online scheduling policy, where the virtual power queue is used to tackle the power constraint. In NOMA systems, our CMDP-based approach achieves a better performance over the Lyapunov approach by taking advantage of the distribution information.

Coordinated multipoint (CoMP) has received considerable attention as a promising technology to improve the transmission rates and spectral efficiency of cell edge users for future networks. How to design coordinated scheduling through a cross-layer approach is challenging in CoMP systems. In this paper, a pilot-efficient scheduling policy is presented in CoMP systems with large-scale antennas. Our policy selects users to access the spectrum and allocates the power to users based on the queue state information (QSI) and channel state information (CSI). Based on the Lyapunov optimization, the cross-layer scheduling problem can be modeled as a combinatorial optimization problem, which is nontrivial. To solve this problem, we decouple the combinatorial optimization problem as a user selection problem and a power allocation problem. Then a low-complexity iteration algorithm is proposed to solve the combinatorial optimization problem. Simulation results demonstrate that our presented policy has better performance over traditional methods. Moreover, by comparing to the exhaustive search algorithm, the performance of our proposed policy is similar to that of an optimal policy.

Non-orthogonal multiple access (NOMA) is recognized as one of the promising techniques in wireless communications. How to reduce the delay violation probability in NOMA systems is a challenging issue. In this paper, a cross-layer scheduling scheme is presented for NOMA systems. The scheduling scheme is designed to jointly determine the scheduling in the network layer and superposition coding process in the physical layer. In order to find the optimal scheduling scheme, we model the queue states of the users as a Markov chain, based on which the delay violation probability and the average power consumption can be analyzed. Then, we minimize the delay violation probability given constraint on average power consumption by formulating and solving a cross-layer optimization problem. We convert the optimization problem into an equivalent linear programming problem via variable substitution, which allows us to obtain the optimal delay-power tradeoff as well as the optimal scheduling policy. One of the optimal superposition coding policies can be determined directly, which can significantly reduce the computational complexity of the linear programming. Theoretical analyses and simulation results show that our approach achieves a better performance over the Longer Queue Highest Possible Rate (LQHPR) policy.