Accurate traffic flow prediction is crucial to alleviating traffic congestion, optimizing road resource allocation, and enhancing public travel experience. However, the complex and dynamic nature of spatial-temporal heterogeneity in traffic flow poses challenges to this task. Existing methods often neglect the interactions between spatial-temporal correlations and struggle to capture dynamic traffic patterns across various perspectives. To address these issues, we propose a Spatial-Temporal Multi-Scale Interactive Graph Convolutional Network (STMIGCN) for traffic flow prediction. In this work, we employ a downsampling-based interactive learning framework to model traffic spatial-temporal patterns from a global-to-local perspective. This framework partitions and extracts spatial-temporal features at different scales, employing an interactive information passing and feedback mechanism to deeply mine latent spatial-temporal correlations. To address the static limitations of adaptive graphs, STMIGCN introduces a multi-graph fusion convolution method that incorporates dynamic features in spatial modeling, thereby enhancing the capture of dynamic spatial relationships between nodes. We perform comprehensive experiments on two real-world datasets and evaluate STMIGCN’s performance against ten baseline models. The results indicate that STMIGCN demonstrates more advanced performance.

Key-value caches, represented by Mem cached, play a critical role in data centers. Its efficacy can significantly impact users' perceived service time and back-end systems' workloads. A central issue in the in-memory cache's management is memory allocation, or how the limited space is distributed for storing key-value items of various sizes. When a cache is full, the allocation issue is how to conduct replacement operations on items of different sizes. To effectively address the issue, a practitioner must simultaneously consider three factors, which are access locality, item size, and miss penalty. Existing designs consider only one or two of the first two factors, and pay little attention on miss penalty. This inadequacy can substantially compromise utilization of cache space and request service time. In this paper we propose a Penalty Aware Memory Allocation scheme (PAMA) that takes all three factors into account. While the three different factors cannot be directly compared to each other in a quantitative manner, PAMA uses their impacts on service time to determine where a unit of memory space should be (de)allocated. The impacts are quantified as the decrease (or increase) of service time if a unit of space is allocated (or deal located). PAMA efficiently tracks access pattern and use of memory, and speculatively evaluates the impacts to enable penalty-aware memory allocation for KV caches. Our evaluation with real-world Mem cached workload traces demonstrates that PAMA can significantly reduce request service time compared to other representative KV cache management schemes.

The objective of maintaining a high efficiency for a shared storage system often has to be compromised with the enforcement of Service-level Agreement (SLA) on quality of service (QoS). From the perspective of I/O scheduling, I/O request service order optimized for disk efficiency can be substantially different from the order required for meeting QoS requirements. When QoS takes priority, the storage system has to serve requests with a sub-optimal efficiency. In this paper, we propose to relate QoS requirements specified for I/O requests to users' experiences. By assuming that only user-observable QoS is necessary for the system to fulfill, we relax QoS requirements on the storage system as long as such a relaxation is not noticeable to users. The relaxation produces a leeway critical for the I/O scheduler to improve disk efficiency. We implement a prototype system, named as intelliQoS, as a proof of concept. In the system the scheduler is allowed to schedule requests in its preferred order as long as the user does not sense any performance degradation through the outputs from the application. In this way the user can still experience the same service quality as required although individual requests' latency requirements can be missed for higher storage efficiency. Our experiments on Xen virtual machines (VMs) show that intelliQoS significantly improves system efficiency by up to 80% without violating user-observable QoS requirements.

Prefetching is an important technique for improving effective hard disk performance. A prefetcher seeks to accurately predict which data will be requested and load it ahead of the arrival of the corresponding requests. Current disk prefetch policies in major operating systems track access patterns at the level of file abstraction. While this is useful for exploiting application-level access patterns, for two reasons file-level prefetching cannot realize the full performance improvements achievable by prefetching. First, certain prefetch opportunities can only be detected by knowing the data layout on disk, such as the contiguous layout of file metadata or data from multiple files. Second, nonsequential access of disk data (requiring disk head movement) is much slower than sequential access, and the performance penalty for mis-prefetching a randomly located block, relative to that of a sequential block, is correspondingly greater.

Challenged by serious power and thermal constraints and limited by available instruction-level parallelism, processor designs have evolved to multi-core architectures. These architectures, many augmented with native simultaneous multithreading, are driving software developers to use multithreaded programs to exploit thread-level parallelism. While multithreading is well known to introduce concerns of data dependency and CPU load balance, less known is that the uncertainty of relative progress of thread execution can cause patterns of I/O requests, issued by different threads, to be effectively random and so significantly degrade hard-disk efficiency. This effect can severely offset the performance gains from parallel execution, especially for I/O-intensive programs. Retaining the benefits of multithreading while not losing I/O efficiency is an urgent and challenging problem. We propose a user-level scheme, iHarmonizer, to streamline the servicing of I/O requests from multiple threads in the Open MP programs. Specifically, we use the compiler to insert code into Open MP programs so that data usage can be transmitted at run time to a supporting run-time library that prefetches data in a disk friendly way and coordinates threads' execution according to the availability of their requested data. Transparent to the programmer, iHarmonizer makes a multithreaded program I/O efficient while maintaining the benefits of parallelism. Our experiments show that iHarmonizer can significantly speed up the execution of a representative set of I/O-intensive scientific benchmarks.

Key-value stores are a vital component in many scale-out enterprises, including social networks, online retail, and risk analysis. Accordingly, they are receiving increased attention from the research community in an effort to improve their performance, scalability, reliability, cost, and power consumption. To be effective, such efforts require a detailed understanding of realistic key-value workloads. And yet little is known about these workloads outside of the companies that operate them. This paper aims to address this gap.

Currently the QoS requirements for disk-based storage systems are usually presented in the form of service-level agreement (SLA) to bound I/O measures such as latency and throughput of I/O requests. However, SLA is not an effective performance interface for users to specify their required I/O service quality for two major reasons. First, for users, it is difficult to determine appropriate latency and throughput bounds to ensure their application performance without resource over-provisioning. Second, for storage system administrators, it is a challenge to estimate a user's real resource demand because the specified SLA measures are not consistently correlated with the user's resource demand. This makes resource provisioning and scheduling less informative and could greatly reduce system efficiency. We propose the concept of reference storage system (RSS), which can be a storage system chosen by users and whose performance can be measured off-line and mimicked on-line, as a performance interface between applications and storage servers. By designating an RSS to represent I/O performance requirement, a user can expect the performance received from a shared storage server servicing his I/O workload is not worse than the performance received from the RSS servicing the same workload. The storage system is responsible for implementing the RSS interface. The key enabling techniques are a machine learning model that derives request-specific performance requirements and an RSS-centric scheduling that efficiently allocates resource among requests from different users. The proposed scheme, named as YouChoose, supports the user-chosen performance interface through efficiently implementing and migrating virtual storage devices in a host storage system. Our evaluation based on trace-driven simulations shows that YouChoose can precisely implement the RSS performance interface, achieve a strong performance assurance and isolation, and improve the efficiency of a consolidated storage system consisting of different types of storage devices.

Currently the QoS requirements for storage systems are usually presented in the form of service-level agreement (SLA) to bound I/O measures such as latency and throughput of I/O requests. However, SLA is not an effective performance interface for users to specify their required I/O service quality for two major reasons. First, for users it is difficult to determine appropriate latency and throughput bounds to ensure their required application performance without resource over-provisioning. Second, for storage system administrators it is a challenge to estimate a user’s real resource demand because the specified SLA measures are not consistently correlated with the user’s resource demand. This makes resource provisioning and scheduling less informative and can greatly reduce system efficiency. We propose the concept of reference storage system (RSS), which can be a storage system chosen by users and whose performance can be measured offline and mimicked online, as a performance interface between applications and storage servers. By designating an RSS to represent I/O performance requirement, a user can expect the performance received from a shared storage server servicing his I/O workload is not worse than the performance received from the RSS servicing the same workload. The storage system is responsible for implementing the RSS interface. The key enabling techniques are a machine learning model that derives request-specific performance requirements and an RSS-centric scheduling that efficiently allocates resource among requests from different users. The proposed scheme, named as YouChoose , supports the user-chosen performance interface through efficiently implementing and migrating virtual storage devices in a host storage system. Our evaluation based on trace-driven simulations shows that YouChoose can precisely implement the RSS performance interface, achieve a strong performance assurance and isolation, and improve the efficiency of a consolidated storage system consisting of different types of storage devices.