In modern high-performance distributed systems, ensuring real-time data consistency and seamless service continuity under massive concurrent updates presents significant engineering challenges. This paper introduces a scalable and efficient architecture for distributed data management, enabling fine-grained incremental updates, low-overhead synchronization, and consistent service transitions. The approach leverages metadata-driven version tracking to isolate update deltas and reduce storage and network overhead. A feedback-regulated multicast protocol ensures efficient data dissemination across nodes without compromising consistency. Additionally, the system employs a layered execution model that allows services to switch versions at runtime with minimal impact on availability. Experiments conducted in a production-simulated environment demonstrate that the proposed architecture improves compression throughput by over 30% and keeps incremental backup size under 5% of the full dataset. The framework can boost multicast synchronization performance by more than 16× over TCP, and reduce update latency under high concurrency by 44%. The framework is applicable to latency-critical, data-intensive domains such as workflow execution, policy-driven systems, and transaction-heavy platforms.

As microservices architectures become increasingly prevalent, they introduce significant operational challenges due to the complexities in service interactions and fault propagation. These architectures often conceal the origins of faults due to intricate inter-service communications, making fault localization both critical and challenging. Addressing these difficulties, this paper introduces a novel fault localization method that leverages synergies between domain prior knowledge, ensemble learning, and graph-based modeling. Our approach models microservices as a graph, with services as nodes and their interactions as edges, illuminating complex dependencies and enhancing the depth of data analysis. The method integrates expert knowledge with a unique blend of multi-class decision trees and strategy models derived from a knowledge base, enabling effective de-tection of diverse patterns and anomalies. Additionally, a meta-learner refines the outputs from base models using a weighted decision-making process, significantly improving the accuracy and robustness of fault detection. Compared to traditional models, including graph neural networks, our approach sub-stantially reduces model complexity and enhances adaptability to evolving service patterns. It demonstrates superior scalability and real-time processing capabilities, offering a robust solution to the challenges of fault localization in dynamic microservice environments.

Adaptive video streaming is very important for delivering high-quality video content of MOOCs (Massive Online Open Courses) to online learners because they often have Internet connections with different levels of bandwidth. Although DASH (Dynamic adaptive streaming over HTTP) is widely accepted as a viable streaming technology to implement scalable Internet streaming using HTTP transport, few research efforts have been made to investigate how to apply this relatively new technology on improving the quality of experience (QoE) of MOOC video streaming. This paper proposes a DASH scheme for MOOC video streaming (MOOC-DASH) to improve QoE of DASH-based MOOC. This scheme consists of content-aware ROI-based video encoding for MOOCs video and bitrate selection algorithm to provide a high-quality and smooth video streaming service to online learners. Experimental results demonstrate that it can effectively reduce the bandwidth of the video content and improve QoE of MOOC Streaming.

Massive amount scientific data often need to be visualized in ultra-large images for scientific discovery. Although ultra-high resolution tiled-display environments have been widely used, there still lacks of proper image viewers that can display ultra-large images with billions of pixels in tiled-display environments. To address the problem, we propose Magic View, an optimized ultra-large scientific image viewer for SAGE tiled-display environment. It can achieve real-time interactive performance in viewing images with billions of pixels. Our experiments show that the performance of Magic View are at lease 8x better than Juxta View, another ultra-large image viewer for SAGE.

Performance of state-of-the-art image retrieval systems has been improved significantly using bag-of-words approaches. After represented by visual words quantized from local features, images can be indexed and retrieved using scalable textual retrieval approaches. However, there exist at least two issues unsolved, especially for search by mobile images with large variations: (1) the loss of features discriminative power due to quantization; and (2) the underuse of spatial relationships among visual words. To address both issues, considering properties of mobile images, this paper presents a novel method coupling visual and spatial information consistently: to improve discriminative power, features of the query image are first grouped using both matched visual features and their spatial relationships; Then grouped features are softly matched to alleviate quantization loss. Experiments on both UKBench database and a collected database with more than one million images show that the proposed method achieves 10% improvement over the approach with a vocabulary tree and bundled feature method.

Real-time video communication such as Internet video conferencing is often afflicted by packet loss over the network. To improve the quality of video, error protection schemes have been introduced based on FMO in H.264 whose encoding efficiency is unacceptable. This paper presents a novel region of interest (ROI) protection scheme that can accurately extract ROI area using facial recognition and greatly speedup video encoding based on feedback using x264 codec implementation. In this scheme, the video receiver uses a packet loss prediction model to predict whether to send feedback to the video sender that dynamically adjust the ROI protecting scheme. Experiments prove that the quality of the ROI area can be effectively improved by the scheme whose encoding performance increases by 50 times compared with FMO based algorithms.

This paper concentrates on speech duration distributions that are usually invariant to noises and proposes a noise-robust and real-time voice activity detector (VAD) using the hidden semi-Markov model (HSMM) to explicitly model state durations. Motivated by statistical observations and tests on TIMIT and the IEEE sentence database, we use Weibull distributions to model state durations approximately and estimate their parameters by maximum likelihood estimators. The final VAD decision is made according to the likelihood ratio test (LRT) incorporating state prior knowledge and modified forward variables. An efficient way that recursively calculates modified forward variables is devised and a dynamic adjustment scheme is used to update parameters. Experiments on noisy speech data show that the proposed method performs more robustly and accurately than the standard ITU-T G.729B VAD and AMR2.

This paper introduces a new approach to build application layer multicast overlay - multiple shared trees. Multiple shared trees' approach makes tradeoffs between traditional source-based trees and single-shared tree, and between transmission efficiency and protocol overheads. Based on this, we propose two protocols to build ALM overlay among end users and media-forwarding-gateways respectively. The latter references the design thought of aggregated multicast to share the multicast trees among groups.