Log parsing is indispensable for system maintenance, converting unstructured log data into structured formats (log templates) for further log compression and anomaly detection. The effectiveness of log parsing relies on the efficiency of two key processes: template extraction and log matching. Traditional methods, however, suffer from slow pairwise comparisons for template extraction and the tedious, non-scalable sequential approach for template matching. Our research has uncovered two opportunities for optimization based on two log locality characteristics: logs from the same template tend to cluster sequentially, and there is a limited variety of templates used within given timeframes. To exploit these opportunities, we propose the Multi-Logs Template Extraction (MLTE)-Cache framework. MLTE-Cache leverages the MLTE algorithm to enhance the efficiency of template extraction by grouping similar logs and processing them in batch mode. Furthermore, the framework utilizes a cache-assisted proximity matching algorithm to accelerate the log matching procedure. Through comprehensive experiments on open-source datasets, the MLTE-Cache framework has proven highly effective, maintaining a high level of accuracy while delivering a 37% improvement in efficiency.

Network virtualization offers a promising solution for the rapid deployment of customized services by enabling the creation of multiple Virtual Networks (VNs) on a shared physical infrastructure. While numerous VN embedding schemes have been developed to efficiently allocate substrate resources for VN requests, limited research has addressed the implementation of robust backup mechanisms to ensure resilience against physical network failures. In virtualized environments, the failure of a single physical server can disrupt multiple virtual nodes, leading to significant service outages. To prevent such disruptions, it is essential to preallocate backup physical nodes and reserve link bandwidth to facilitate the migration or remapping of VN requests in the event of any failure. Ensuring the survivability of embedded VNs in a resource-efficient manner is thus imperative. In this work, we introduce a novel approach that utilizes star-based graph decomposition combined with dynamic programming to optimize the mapping of virtual nodes and edges to physical network resources. Simulation results demonstrate that our proposed scheme achieves a higher acceptance ratio and more efficient utilization of physical resources compared to existing backup strategies.

Bloom Filters (BFs) are compact and probabilistic data structures designed for efficient set membership queries. They offer high query and storage efficiency, making them particularly useful in network and distributed systems. However, the scalability of BFs in accommodating “big data” is limited by increased false positive rates, inflexible hash functions, and inefficient matching with dynamic datasets. To address these limitations, we introduce the Extensible Bloom Filter (EBF), which incorporates a flexible expansion mechanism and an adaptive hash function generation scheme. The EBF design features a set of BF vectors that expand according to the rate of incoming data, with each vector sized to suit the characteristics of the data. Adaptive hash functions, derived from common base matrices, streamline the process by leveraging strong inter-hash relationships. This reduces overhead and simplifies queries across multiple BF vector sizes. Performance evaluations have shown that the EBF consistently achieves a low false positive rate and minimal query time, even amid dynamic data arrivals and large data sets. With its extensibility and adaptability, the EBF provides a robust solution for applications requiring dynamic set representations with stringent accuracy requirements. It enhances the capabilities of network and distributed systems, making them more efficient in handling complex data scenarios.