Neural Network Hot Spot Prediction Algorithm for Shared Web Caching System

Access to the internet and WWW is growing extensively, which results in heavy network traffic. To reduce the network traffic proxy server is used. Proxy server reduces the load of server. If the cache replacement algorithm of proxy server's cache is efficient then proxy server will be helpful to reduce the network traffic in more efficient manner. In this paper we are considering proxy server cache to be Level 1 (L1) cache and storage cache of proxy server to be Level 2 (L2) cache. For collecting the real trace CC proxy server is used in an organization. Log of proxy server gives the information of various URLs accessed by various clients with time. For performing experiments various URLs were given a numeric identity. This paper proposes an efficient replacement algorithm on L1 for proxy server. The replacement algorithms taken into consideration are Least Recently Used (LRU), Least Frequently Used (LFU), First In First Out (FIFO). Access Pattern of L1 and L2 are different as if the desired page is not on L1 then it accesses to L2. Thus L1 is having better temporal locality than L2. Thus the replacement algorithm which is giving efficient results for L1 may not be suitable for L2. This paper also proposes a preeminent pair of replacement algorithms for L1 and L2 cache for proxy server.

Transport layer security is an important security protocol, which is used to protect end-to-end communication. However, limitation occurs when it is applied to content delivery networks, in which the proxy server rather than the origin server provides service to the client. Under such circumstances, the proxy server serves as a third party and the client is not able to authenticate. This paper discusses the authentication problem for the proxy server. Afterwards, a delegation token-based method is proposed to authenticate the proxy server, with multi-level proxy servers being taken into consideration. Furthermore, a client-based cache strategy is employed to improve the proposed method in terms of time consumption. Then the security of the method is also analysed. Experimental results demonstrate the effectiveness of our method. Moreover, with client-based cache strategy, the authentication process can be accomplished much more efficiently, with a 15.63% decrease in terms of connection time.

To meet the growing energy management needs of the rail transit system, this study explores the configuration and scheduling strategies of rail transit energy storage. An innovative prediction-allocation-dispatch co-optimization framework based on the deep fusion of the Beetle Antennae search algorithm and back-propagation neural network is proposed. This framework solves the problem of the co-optimization of renewable energy output uncertainty and the economy of energy storage. The study adopts a Beetle Antennae-optimized backpropagation neural network prediction algorithm to improve the prediction accuracy of wind and solar power generation. A hybrid energy storage configuration system is designed by comprehensively considering the energy storage life and cost, and a multi-time scale energy scheduling optimization scheme is proposed. The experiment showed that the optimization prediction algorithm could significantly improve prediction accuracy and had superior iterative optimization effects. Compared with traditional prediction algorithms, this optimization algorithm reduced the mean square error of photovoltaic and wind power prediction by 41.11% and 34.02% respectively and reduced the time consumption by 29.95% and 33.21% compared to other algorithms. Meanwhile, the hybrid model had significant advantages in energy storage costs and net benefits, with a total scheduling cost reduction of 17.95% compared to other energy storage systems and a response speed improvement of 13.98%. This indicates that the proposed optimization algorithm and configuration model can enhance the energy utilization function and scheduling flexibility of the rail transit system, supporting the sustainable development of rail transit.

Hot spots are residues comprising only a small fraction of interfaces yet accounting for the majority of the binding energy. These residues are critical in understanding the principles of protein interactions. Experimental studies like alanine scanning mutagenesis require significant effort; therefore, there is a need for computational methods to predict hot spots in protein interfaces. We present a new intuitive efficient method to determine computational hot spots based on conservation (C), solvent accessibility [accessible surface area (ASA)] and statistical pairwise residue potentials (PP) of the interface residues. Combination of these features is examined in a comprehensive way to study their effect in hot spot detection. The predicted hot spots are observed to match with the experimental hot spots with an accuracy of 70% and a precision of 64% in Alanine Scanning Energetics Database (ASEdb), and accuracy of 70% and a precision of 73% in Binding Interface Database (BID). Several machine learning methods are also applied to predict hot spots. Performance of our empirical approach exceeds learning-based methods and other existing hot spot prediction methods. Residue occlusion from solvent in the complexes and pairwise potentials are found to be the main discriminative features in hot spot prediction. Our empirical method is a simple approach in hot spot prediction yet with its high accuracy and computational effectiveness. We believe that this method provides insights for the researchers working on characterization of protein binding sites and design of specific therapeutic agents for protein interactions. The list of training and test sets are available as Supplementary Data at http://prism.ccbb.ku.edu.tr/hotpoint/supplement.doc. Supplementary data are available at Bioinformatics online.

Traditionally, file transfer protocol (FTP) servers are major archive providers and users apply the Archie server to locate FTP archives. With the extreme popularity of the WWW, Web servers are now important archive providers and users download archives via HTTP. To reduce HTTP traffic, proxy cache servers are deployed on the Internet. However, we find the hit rate of archives in cache servers is quite low. This study proposes a combination of caching and better searching mechanisms to alleviate the problem. We enable a proxy server to automatically collect WWW and FTP archives from its cache, organize them in the form of an FTP directory, and then offer the directory list to the Archie. Accordingly, users can find archives on WWW and FTP servers through the Archie, and they can directly download archives from the proxy server, thus improving the reuse of cached archives. A system was implemented and operated in a real environment to evaluate the approach and results are discussed.

Neural network prediction algorithms in tool production

Caching web pages at proxies and in web servers' memories can greatly enhance performance. Proxy caching is known to reduce network and both proxy and server caching can significantly decrease latency. Web caching problems have different properties than traditional operating systems caching, and replacement can benefit by recognizing and exploiting these differences. We address two aspects of the predictability of traffic patterns: the overall experienced by large proxy and web servers, and the distinct access patterns of individual pages. We formalize the notion of cache load under various replacement policies, includingLRU andLFU, and demonstrate that the trace of a large proxy server exhibits regular load. Predictable allows for improved design, analysis, and experimental evaluation of replacement policies. We provide a simple and (near) optimal replacement policy when each page request has an associated distribution function on the next request time of the page. Without the predictable assumption, no such online policy is possible and it is known that even obtaining an offline optimum is hard. For experiments, predictable enables comparing and evaluating replacement policies using partial traces, containing requests made to only a subset of the pages. Our results are based on considering a simpler caching model which we call the interval caching model .W e relate traditional and interval caching policies under predictable load, and derive (near)-optimal replacement policies from their optimal interval caching counterparts.

The World Wide Web suffers from scaling and reliability problems due to overloaded and congested proxy servers. Caching at local proxy serves help, but cannot satisfy more than a third to half of requests; more requests are still sent to original remote servers. In this paper we have developed an algorithm for Distributed Web Cache, which incorporates cooperation among proxy servers of one cluster. This algorithm uses Distributed Web Cache concepts along with static hierarchies with geographical based clusters of level one proxy server with dynamic mechanism of proxy server during the congestion of one cluster. Congestion and scalability problems are being dealt by clustering concept used in our approach. This results in higher ratio of caches, with lesser latency delay for requested pages. This algorithm alsi guarantees data consistency between the original server objects and the proxy cache objects.KeywordsDistributed Web CachingClusteringProxy ServerLatencyHit Ratio

An Internet flash crowd (also known as hot spots) is a phenomenon that results from a sudden, unpredicted increase in an on-line object's popularity. Currently, there is no efficient means within the Internet to deliver Web objects scalably under hot spot conditions to all clients that desire the object. We present peer-to-peer (P2P) randomized overlays to obviate flash-crowd symptoms (PROOFS), a simple, lightweight, P2P approach that uses randomized overlay construction and randomized, scoped searches to locate and deliver objects efficiently under heavy demand to all users that desire them. We evaluate PROOFS' robustness in environments in which clients join and leave the P2P network, as well as in environments in which clients are not always fully cooperative. Through a mix of simulation and prototype experimentation in the Internet, we show that randomized approaches like PROOFS should effectively relieve flash crowd symptoms in dynamic, limited-participation environments.

Hot spots are the subset of interface residues that account for most of the binding free energy, and they play essential roles in the stability of protein binding. Effectively identifying which specific interface residues of protein–protein complexes form the hot spots is critical for understanding the principles of protein interactions, and it has broad application prospects in protein design and drug development. Experimental methods like alanine scanning mutagenesis are labor-intensive and time-consuming. At present, the experimentally measured hot spots are very limited. Hence, the use of computational approaches to predicting hot spots is becoming increasingly important. Here, we describe the basic concepts and recent advances of machine learning applications in inferring the protein–protein interaction hot spots, and assess the performance of widely used features, machine learning algorithms, and existing state-of-the-art approaches. We also discuss the challenges and future directions in the prediction of hot spots.

The KFC Server is a web-based implementation of the KFC (Knowledge-based FADE and Contacts) model—a machine learning approach for the prediction of binding hot spots, or the subset of residues that account for most of a protein interface's; binding free energy. The server facilitates the automated analysis of a user submitted protein–protein or protein–DNA interface and the visualization of its hot spot predictions. For each residue in the interface, the KFC Server characterizes its local structural environment, compares that environment to the environments of experimentally determined hot spots and predicts if the interface residue is a hot spot. After the computational analysis, the user can visualize the results using an interactive job viewer able to quickly highlight predicted hot spots and surrounding structural features within the protein structure. The KFC Server is accessible at http://kfc.mitchell-lab.org.

The General Packet Radio Service (GPRS) is being deployed by GSM network operators world-wide, and promises to offer users always-on data access at bandwidths comparable to that of conventional fixed-line telephone modems. Unfortunately, many users have found the reality to be rather different, experiencing very disappointing performance when, for example, browsing the web over GPRS.In this paper we investigate what causes the HTTP protocol and its underlying transport TCP to underperform in a GPRS environment. We examine why certain GPRS network characteristics interact badly with TCP to yield problems such as: link under-utilization for short-lived flows, excess queueing for long-lived flows, ACK compression, poor loss recovery, and gross unfairness between competing flows. We also show that many web browsers tend to be overly aggressive, and by opening too many simultaneous TCP connections can aggravate matters.We present the design and implementation of GPRSWeb - a mobile HTTP proxy system that mitigates many of the performance problems with a simple software update to a GPRS mobile device. The update is a 'client proxy' that sits in the mobile device, and communicates with a 'server proxy' located at the other end of the GPRS link close to the wired-wireless border. The dual proxy architecture collectively implements a number of key enhancements - an aggressive caching scheme that employs content-based hash keying to improve hit rates for dynamic content, a preemptive push of web page support resources to mobile clients, resource adaptation to suit client capabilities, delta encoded data transfers, DNS lookup migration, and a UDP-based reliable transport protocol that is specifically optimized for use over GPRS. We show that these enhancements result in significant improvement in overall WWW performance over GPRS.

BackgroundIdentification of hot spots in protein-DNA interfaces provides crucial information for the research on protein-DNA interaction and drug design. As experimental methods for determining hot spots are time-consuming, labor-intensive and expensive, there is a need for developing reliable computational method to predict hot spots on a large scale.ResultsHere, we proposed a new method named sxPDH based on supervised isometric feature mapping (S-ISOMAP) and extreme gradient boosting (XGBoost) to predict hot spots in protein-DNA complexes. We obtained 114 features from a combination of the protein sequence, structure, network and solvent accessible information, and systematically assessed various feature selection methods and feature dimensionality reduction methods based on manifold learning. The results show that the S-ISOMAP method is superior to other feature selection or manifold learning methods. XGBoost was then used to develop hot spots prediction model sxPDH based on the three dimensionality-reduced features obtained from S-ISOMAP.ConclusionOur method sxPDH boosts prediction performance using S-ISOMAP and XGBoost. The AUC of the model is 0.773, and the F1 score is 0.713. Experimental results on benchmark dataset indicate that sxPDH can achieve generally better performance in predicting hot spots compared to the state-of-the-art methods.

Wild-type green fluorescent protein (GFP) has a weak dimerization tendency at higher concentrations. Currently there are still a limited number of known mutated residues for the dimer interface breaking. The crystal structure of the dimer provides vital background on neighbouring residue interactions. Here well-established hot spot prediction methods, such as HotPoint, KFC2, PredHS and Robetta, were adopted to predict residues critical for dimer stabilization, i.e. hot spots. Among 17 residues appeared on the dimer interface of GFP, 10 residues were predicted as hot spots by at least one method. Some of these computational hot spots were verified with existing site-directed mutagenesis results, and the remains of the predicted residues can be candidates for further validation. Furthermore, the structural feature such as relative accessibility was found significantly different between computational hot spots and non-hot spots.

Aiming at the problems of low accuracy and poor prediction effect in traditional e-commerce network users' hot spot prediction, this paper proposes an e-commerce network users' hot spot prediction method based on improved K-nearest neighbour algorithm. According to the users' hotspot prediction impact indicator, the K-nearest neighbour algorithm is improved in pattern matching process. The key point method is used to remove the noise interference of the original time series. The dynamic time warping algorithm is used to measure the similarity of time series of users' hotspot. The distance weight and trend coefficient are introduced according to the difference of users' hotspot time series to deduce the future users' hotspot and realise hotspot prediction of e-commerce network users. Experimental results show that the method in this paper greatly reduces the deviation of prediction results, which fully shows that the method has better prediction effect.