摘要:
Apparatus and methods for identifying traffic patterns to web sites based on templates that characterize the arrival of traffic to the web sites are provided. Based on these templates, determinations are made as to which web sites should be co-located so as to optimize resource allocation. Specifically, web sites whose templates are complimentary, i.e. a first web site having a peak in arrival traffic at time t1 and a second web site that has a trough in arrival traffic at time t1, are designated as being candidates for co-location. In addition, the present invention uses the templates identified for the traffic patterns of web sites to determine thresholds for offloading traffic to other servers. These thresholds include a first threshold at which offloading should be performed, a second threshold that takes into consideration the lead time needed to begin offloading, and a third threshold that takes into consideration a lag time needed to stop all offloading of traffic to the other servers.
摘要:
Apparatus and methods for classifying web sites are provided. With the apparatus and methods, traffic data is obtained for a plurality of web sites. This patterns, or templates, for each web site are generated based on this traffic data and the patterns are clustered into classes of web sites using a clustering algorithm. The clusters, or classes, are then profiled to generate a template for each class. The template for each class is generated by first shifting the patterns for each web site that is part of the class to compensate for effects like time zone differences, if any, and then identifying a pattern that is most similar to all of the patterns in the class. Once the template for each class is generated, this template is then used with traffic data from a new web site to classify the new web site into one of the existing classes. In other words, when traffic data for a new web site is received, a pattern for the traffic data of the new web site is generated and compared to the templates for the various classes. If a matching class template is identified, the new web site is classified into the corresponding class. If the pattern for the new web site does not match any of the existing templates, a new template and class may be generated based on the pattern for the new web site.
摘要:
Identifying traffic patterns to web sites based on templates that characterize the arrival of traffic to the web sites is provided. Based on these templates, determinations are made as to which web sites should be co-located so as to optimize resource allocation. Web sites whose templates are complimentary, i.e. a first web site having a peak in arrival traffic at time t1 and a second web site that has a trough in arrival traffic at time t1, are designated as being candidates for co-location. In addition, the templates identified for the traffic patterns of web sites are used to determine thresholds for offloading traffic to other servers. These thresholds include a first threshold at which offloading should be performed, a second threshold that takes into consideration the lead time needed to begin offloading, and a third threshold that takes into consideration a lag time needed to stop offloading of traffic.
摘要:
Apparatus and methods for identifying traffic patterns to web sites based on templates that characterize the arrival of traffic to the web sites are provided. Based on these templates, determinations are made as to which web sites should be co-located so as to optimize resource allocation. Specifically, web sites whose templates are complimentary, i.e. a first web site having a peak in arrival traffic at time t1 and a second web site that has a trough in arrival traffic at time t1, are designated as being candidates for co-location. In addition, the present invention uses the templates identified for the traffic patterns of web sites to determine thresholds for offloading traffic to other servers. These thresholds include a first threshold at which offloading should be performed, a second threshold that takes into consideration the lead time needed to begin offloading, and a third threshold that takes into consideration a lag time needed to stop all offloading of traffic to the other servers.
摘要:
Methods and apparatus operating in a stream processing network perform load shedding and dynamic resource allocation so as to meet a pre-determined utility criterion. Load shedding is envisioned as an admission control problem encompassing source nodes admitting workflows into the stream processing network. A primal-dual approach is used to decompose the admission control and resource allocation problems. The admission control operates as a push-and-pull process with sources pushing workflows into the stream processing network and sinks pulling processed workflows from the network. A virtual queue is maintained at each node to account for both queue backlogs and credits from sinks. Nodes of the stream processing network maintain shadow prices for each of the workflows and share congestion information with neighbor nodes. At each node, resources are devoted to the workflow with the maximum product of downstream pressure and processing rate, where the downstream pressure is defined as the backlog difference between neighbor nodes. The primal-dual controller iteratively adjusts the admission rates and resource allocation using local congestion feedback. The iterative controlling procedure further uses an interior-point method to improve the speed of convergence towards optimal admission and allocation decisions.
摘要:
Techniques are provided for determining a capacity allocation in a multi-tiered computing system. In one aspect of the invention, a technique for designing capacity allocation for a multi-tiered computing system, each tier of the computing system having one or more computing devices (e.g., servers), comprises the following steps/operations. Input parameters relating to capacity allocation are obtained. A capacity allocation is computed based on at least a portion of the input parameters, the capacity allocation being computable such that one or more end-to-end performance measures are substantially satisfied for multiple service classes. The capacity allocation computation may minimize a cost function. The capacity allocation computation may also satisfy one or more mean delay guarantees for the multiple service classes, one or more tail distribution guarantees for the multiple service classes, or both one or more mean delay guarantees and one or more tail distribution guarantees for the multiple service classes.
摘要:
A system and method for resource allocation includes, in a network having nodes and links, injecting units of flow for at least one commodity at a source corresponding to the at least one commodity. At each node, queue heights, associated with the at least one commodity, are balanced for queues associated with each of one or more outgoing paths associated with that node. An amount of commodity flow is pushed across a link toward a sink, where the amount of commodity flow is constrained by a capacity constraint. Flow that reached the sink is absorbed by draining the queues.
摘要:
Method for modeling the performance of an Information Technology system are disclosed. The method includes the steps of receiving performance data of the system; receiving data of transaction flows and system topology; and inferring service demand parameters based on the received data. If closed form expressions are available to characterize the system, an optimization algorithm based on minimum distance between predicted and measured response times and may be used to obtain the performance parameters. Alternatively, a discrete event simulator together with a set of meta-heuristic search methods may be applied to obtain the optimized performance parameters.
摘要:
In one embodiment, the present invention is a method and apparatus for automatic system parameter configuration for performance improvement. One embodiment of the inventive method involves formulating a black box optimization problem, and solving the optimization problem using an enhanced smart hill climbing method. The smart hill climbing method includes both a global and a more precise local search to identify an optimal solution. In one embodiment, one or both of the global and local searches employs a weighted Latin Hypercube Sampling method in combination with importance sampling techniques to yield improved search results.
摘要:
In one embodiment, the present invention is a method and apparatus for automatic system parameter configuration for performance improvement. One embodiment of the inventive method involves formulating a black box optimization problem, and solving the optimization problem using an enhanced smart hill climbing method. The smart hill climbing method includes both a global and a more precise local search to identify an optimal solution. In one embodiment, one or both of the global and local searches employs a weighted Latin Hypercube Sampling method in combination with importance sampling techniques to yield improved search results