For mixed HDD and SSD storage scenarios, Ceph Cache Tier provides a tiered caching feature that separates fast and slow storage pools to manage data objects more efficiently. However, due to the limited total capacity...
详细信息
ISBN:
(数字)9789819708345
ISBN:
(纸本)9789819708338;9789819708345
For mixed HDD and SSD storage scenarios, Ceph Cache Tier provides a tiered caching feature that separates fast and slow storage pools to manage data objects more efficiently. However, due to the limited total capacity of the cache pool, only some data objects can be stored. Performance can be significantly improved when clients focus on accessing hot objects in the cache pool. If a client accesses the cache pool without hitting data, redundant IO operations occur, which increases client access latency and reduces throughput. To improve the hit rate of the Ceph Cache Tier cache pool, this paper proposes a temperature density-based cache replacement algorithm (TDC). The algorithm improves the hit rate of the cache pool by calculating the temperature density of the space consumed by each object and evicting objects with the lowest temperature density, thus evicting objects that contribute less to the hit rate. The algorithm mainly includes object temperature calculation, temperature density calculation and cache replacement policy. Subsequently, we evaluate the TDC algorithm on a real traces dataset using playback workload IO and demonstrate the efficiency of the algorithm in improving the cache hit rate. Finally, we applied the TDC algorithm to a Ceph distributed storage system and verified the performance of the Cache Tier based on the TDC algorithm.
Content Delivery Cloud (CDC) extends Content Delivery Network (CDN) to provide elastic, scalable and low cost services to the customers. For multimedia streaming over CDC, caching the media content onto the edge serve...
详细信息
ISBN:
(纸本)9781467348638
Content Delivery Cloud (CDC) extends Content Delivery Network (CDN) to provide elastic, scalable and low cost services to the customers. For multimedia streaming over CDC, caching the media content onto the edge server from storage cloud is commonly used to minimize the latency of content delivery. It is very important for CDN to balance between the resources being used (storage space, bandwidth, etc) and the performance achieved. Commercial CDNs (such as Akamai, Limelight, Amazon CloudFront) have their proprietary caching algorithms to deal with this issue. In this paper, we propose a method to further improve the efficiency of the caching system for scalable multimedia contents. Specifically, we notice that a scalable multimedia content can be flexibly truncated to lower bit rates on-the-fly based on the available network bandwidth between the edge server to the end users. Therefore, it may not be necessary to cache such a content at its highest quality/rate. Based on this observation, we show that edge server can decide an optimized truncation ratio for the cached scalable multimedia contents to balance between the quality of the media and the resource usage. The proposed optimized truncation algorithm is analyzed and its efficacy in improving the efficiency of the caching system is justified with simulation result.
Tensor completion is the task of completing multi-aspect data represented as a tensor by accurately predicting missing entries in the tensor. It is mainly solved by tensor factorization methods, and among them, Tucker...
详细信息
ISBN:
(纸本)9781538691595
Tensor completion is the task of completing multi-aspect data represented as a tensor by accurately predicting missing entries in the tensor. It is mainly solved by tensor factorization methods, and among them, Tucker factorization has attracted considerable interests due to its powerful ability to learn latent factors and even their interactions. Although several Tucker methods have been developed to reduce the memory and computational complexity, the state-of-the-art method still 1) generates redundant computations and 2) cannot factorize a large tensor that exceeds the size of memory. This paper proposes FTCOM, a fast and scalable Tucker factorization method for tensor completion. FTCOM performs element-wise updates for factor matrices based on coordinate descent, and adopts a novel caching algorithm which stores frequently-required intermediate data. It also uses a tensor file for disk-based data processing and loads only a small part of the tensor at a time into the memory. Experimental results show that FTCOM is much faster and more scalable compared to all other competitors. It significantly shortens the training time of Tucker factorization, especially on real-world tensors, and it can be executed on a billion-scale tensor which is bigger than the memory capacity within a single machine.
Wireless wearable embedded devices dominate the Internet of Things (IoT) due to their ability to provide useful information about the body and its local environment. The constrained resources of low power processors, ...
详细信息
ISBN:
(纸本)9781538619711
Wireless wearable embedded devices dominate the Internet of Things (IoT) due to their ability to provide useful information about the body and its local environment. The constrained resources of low power processors, however, pose a significant challenge to run-time error logging and hence, product reliability. Error logs classify error type and often system state following the occurrence of an error. Traditional error logging algorithms attempt to balance storage and accuracy by selectively overwriting past log entries. Since a specific combination of firmware faults may result in system instability, preserving all error occurrences becomes increasingly beneficial as IOT systems become more complex. In this paper, a novel hash-based error logging algorithm is presented which has both constant insertion time and constant memory while also exhibiting no false negatives and an acceptable false positive error rate. Both theoretical analysis and simulations are used to compare the performance of the hash-based and traditional approaches.
The efficient distribution of stored information has becomea major concern in the Internet. Since the web workloadcharacteristics shows that more than 60% of network traffic iscaused by image documents, how to efficie...
详细信息
The efficient distribution of stored information has becomea major concern in the Internet. Since the web workloadcharacteristics shows that more than 60% of network traffic iscaused by image documents, how to efficiently distributeimage documents from servers to end clients is an importantissue. Proxy cache is an efficient solution to reduce networktraffic. And it has been shown that an image caching method(Graceful caching) based on hierarchical coding formatshowed better performance than conventional cachingschemes in recent years. However, as the capacity of the cacheis limited, how to efficiently allocate the cache memory toachieve a minimum expected delay time is still a problem to beresolved. This paper presents an integrated caching algorithm to dealwith the above problem in the Internet. By analyzing the webrequest distribution of Graceful caching, both replacementand pre-fetching algorithms are proposed. We also show thatour proposal can be carried out based on information readilyavailable in the proxy server and it flexibly adapts itsparameters to the hit rates and access pattern of users'requesting documents in the Graceful caching. we finallyverify the performance of this algorithm by simulations.
Since the traffic of NetNews is increasing, storage of all articles becomes a serious problem from the viewpoint of wasting network bandwidth and the amount of disk usage. In addition, not all incoming articles are re...
详细信息
Today, Video on Demand (VoD) is a digital service on the rise that requires a lot of resources for its implementation. To reduce the costs of running this service, one of the commonly used alternatives is using proxie...
详细信息
Today, Video on Demand (VoD) is a digital service on the rise that requires a lot of resources for its implementation. To reduce the costs of running this service, one of the commonly used alternatives is using proxies that cache the most important portions of the collection in order to meet the demand for this content in place of the primary server of the VoD system. In this context, to improve the efficiency of proxy, we proposed a novel caching algorithm that explores the positioning of the active clients to determine the density of clients inside a time window existing in front of each video chunk. By caching the video chunks with the greater density in front of them, the algorithm is able to achieve high performance, in terms of the hit ratio for the requests received by the proxy, during periods of high workload. To better evaluate our approach, we compare it with others of similar nature, using both traditional metrics like hit rate, as well as physical metrics, such as the use of processing resources. The results show that the new algorithm exploits the processing bandwidth available in the underlying architecture of the proxy for obtaining a larger hit rate in comparison to the other algorithms used in the comparative analysis. Finally, to dispose of the necessary tools to perform this analysis, we produced another important contribution in this work: the implementation of a VoD proxy simulator that, to the best of our knowledge, is the first one to enable the evaluation of the hardware used to implement this application. ...
暂无评论