Browse Prior Art Database

System And Method For Autonomic Consolidation Of IT Resources In A Cloud Environment By Mining For Complimentary Temporal Workload Characteristics

IP.com Disclosure Number: IPCOM000206732D
Publication Date: 2011-May-05
Document File: 6 page(s) / 126K

Publishing Venue

The IP.com Prior Art Database

Abstract

Cloud computing is gaining widespread acceptance. Many public cloud providers have sprung up in the market to take advantage of these growing markets. Several large enterprises have also moved to the private cloud model. Apart from the basic reasons such as reliability, scalability, time to market etc..better ROI (Return On Investment) and cost reduction are the major key ingredients for survival and sustenance of this cloud model. Irrespective of cloud offerings being private or public, effective and optimized use of IT resources can not only provide better throughput but also free up unused resources for new applications. In other words, optimized allocation of workloads to resources can free up resources and avoid any performance bottlenecks. Since the consumers just define the SLA with provider(that can potentially create over-provisioning), it is the responsibility of the provider to find out complementary workloads and consolidate them together to create an optimized resource usage setup. Applications are deployed in public or private cloud environments based on the defined SLAs. Due to the dynamic nature of the environment (change in the environment and change in application characteristics), often times either over-provisioning or resource-bottlenecks are noticed. Efficient and optimized resource allocation can not only eliminate resource bottlenecks but also free up resources that can be used to host other applications making the provider more profitable. There are consolidation frameworks available in each individual layers of the IT stack (such as servers, virtual machines etc..) and most of them consider the maximum or average usage to recommend consolidation or try to load-balance the resources. Our framework on the other hand takes the whole stack into consideration along with the provider's priorities to create a complementary consolidation clusters that have temporal affinity across multiple dimensions so that a provider can compare-contrast.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 38% of the total text.

Page 01 of 6

System And Method For Autonomic Consolidation Of IT Resources In A Cloud Environment By Mining For Complimentary Temporal Workload Characteristics

Cloud computing is gaining widespread acceptance. Many public cloud providers have sprung up in the market to take advantage of these growing markets. Several large enterprises have also moved to the private cloud model. Apart from the basic reasons such as reliability, scalability,
time to market etc..better ROI (Return On Investment) and cost reduction are the major key ingredients for survival and sustenance of this cloud model.

Irrespective of cloud offerings being private or public, effective and optimized use of IT resources can not only provide better throughput but also free up unused resources for new applications. In other words, optimized allocation of workloads to resources can free up resources and avoid any performance bottlenecks. Since the consumers just define the SLA with provider(that can potentially create over-provisioning), it is the responsibility of the provider to find out complementary workloads and consolidate them together to create an optimized resource usage setup.

Applications are deployed in public or private cloud environments based on the defined SLAs. Due to the dynamic nature of the environment (change in the environment and change in application characteristics), often times either over-provisioning or resource-bottlenecks are noticed. Efficient and optimized resource allocation can not only eliminate resource bottlenecks but also free up resources that can be used to host other applications making the provider more profitable.

There are consolidation frameworks available in each individual layers of the IT stack (such as servers, virtual machines etc..) and most of them consider the maximum or average usage to recommend consolidation or try to load-balance the resources. Our framework on the other hand takes the whole stack into consideration along with the provider's priorities to create a complementary consolidation pairs that have temporal affinity across multiple dimensions so that a provider can compare-contrast.

Our proposed framework mines the resource usage and performance of applications over time to derive an uniform metrics that can be used to create potential consolidation pairs. Pairs are created by taking into consideration of cloud provider's weight assignment to applications/IT resources/etc.... Cloud provider s can then examine the ranked consolidation tuples across multiple dimensions to weigh the benefits and make a decision.

This invention analyzes temporal resource usage statistics (e.g. IOPS for storage system components) across workload profiles, in order to find complimentary patterns that would render two or more workload profiles good candidates for consolidation). These patterns can be temporal and/or spatial (i.e. in terms of the resource the workloads are bound to), for example, whenever one workload reaches its peak res...