OPTIMIS: Improving Cloud Management With Dynamic SLAs



Similar documents
Why Use OPTIMIS? Build and Run Services in the Most Suitable Cloud Venues. May 2013

Cloud Federations in Contrail

The OPTIMIS Project. Optimized Infrastructure Services. Scientific Results

D Self-managed Cloud runtime prototype

How do we define cloud computing?

Document downloaded from: This paper must be cited as:

Requirements and Architecture of a. Cloud Brokerage

Daniel Field, Atos Spain. Towards the European Open Science Cloud, Heidelberg, 20/01/2016

How To Manage Cloud Service Provisioning And Maintenance

Energy Efficiency Embedded Service Lifecycle: Towards an Energy Efficient Cloud Computing Architecture

IaaS Federation. Contrail project. IaaS Federation! Objectives and Challenges! & SLA management in Federations 5/23/11

Telecom Asset Lifecycle Management Tool

CloudLink - The On-Ramp to the Cloud Security, Management and Performance Optimization for Multi-Tenant Private and Public Clouds

Provisioning Server Service Template

Multifaceted Resource Management for Dealing with Heterogeneous Workloads in Virtualized Data Centers

FLEXIANT. Utility Computing on Demand

Scheduling and Monitoring of Internally Structured Services in Cloud Federations

Requirements and Architecture of a. Cloud Brokerage

ActiveVOS Server Architecture. March 2009

QoS Resource Management for Cloud Federations

Service Design, Management and Composition: Service Level Agreements Objectives

A Model for Accomplishing and Managing Dynamic Cloud Federations

Management and Web service Management

Figure 1: Illustration of service management conceptual framework

Work in Progress on Cloud Computing in Myriads Team and Contrail European Project Christine Morin, Inria

WEBAPP PATTERN FOR APACHE TOMCAT - USER GUIDE

Service & Network Management

CompatibleOne Open Source Cloud Broker Architecture Overview

SAP Secure Operations Map. SAP Active Global Support Security Services May 2015

Managing Cloud Computing Risk

Cisco Intercloud Fabric Security Features: Technical Overview

IaaS Cloud Architectures: Virtualized Data Centers to Federated Cloud Infrastructures

A Reference Architecture for Multi-Level SLA Management

Flexiant Cloud Orchestrator with Parallels Cloud Server

SLA Business Management Based on Key Performance Indicators

Privacy-Aware Scheduling for Inter-Organizational Processes

DEFINING GENERIC ARCHITECTURE FOR CLOUD IAAS PROVISIONING MODEL

SLA-BASED RESOURCE MANAGEMENT AND ALLOCATION

SeaClouds Project. Cloud Application Programming Interface. Seamless adaptive multi- cloud management of service- based applications

Implementing a Microsoft SQL Server 2005 Database

Clarity Assurance allows operators to monitor and manage the availability and quality of their network and services

ADDENDUM TO STATE OF MARYLAND PURCHASES ISSUED UNDER STATE CONTRACT NO. 060B

Transforming Service Life Cycle Through Automation with SDN and NFV

Multi-Level SLAs with Dynamic Negotiations for Remote Sensing Data as a Service

Simplify IT. With Cisco Application Centric Infrastructure. Roberto Barrera VERSION May, 2015

Blueprinting Approach in Support of Cloud Computing. Dinh Khoa Nguyen*, Francesco Lelli, Mike P. Papazoglou, and Willem-Jan Van Den Heuvel

Configuration Notes 0215

Top Purchase Considerations for Virtualization Management

IBM Web Server as a Service

Journey to the Private Cloud. Key Enabling Technologies

Managing SOA Security and Operations with SecureSpan

Web Application Hosting Cloud Architecture

MENTER Overview. Prepared by Mark Shayman UMIACS Contract Review Laboratory for Telecommunications Science May 31, 2001

Building an AWS-Compatible Hybrid Cloud with OpenStack

Standardised SLAs: how far can we go? DIHC, Euro-Par 2013, Aachan John Kennedy Intel Labs Europe

SLA Model Terms and Specifications: SLALOM Project Overview and Request for Feedback

Metastorm BPM Interwoven Integration. Process Mapping solutions. Metastorm BPM Interwoven Integration. Introduction. The solution

Environments, Services and Network Management for Green Clouds

Software Define Storage (SDs) and its application to an Openstack Software Defined Infrastructure (SDi) implementation

A Strawman Model. NIST Cloud Computing Reference Architecture and Taxonomy Working Group. January 3, 2011

Article 29 Working Party Issues Opinion on Cloud Computing

SERVICE LEVEL AGREEMENT PROVISIONING AND MONITORING FOR END TO END QOS

Transcription:

1 OPTIMIS: Improving Cloud Management With Dynamic SLAs Salt Lake City, July 18 Wolfgang Ziegler, Fraunhofer Institute SCAI wolfgang.ziegler@scai.fraunhofer.de

OPTIMIS Project IP 5th call June 2010 - May 2013 13 Partner Atos Origin SAE Umeå University The 451 Group Universität Stuttgart National Technical University of Athens Barcelona Supercomputing Center SAP Fraunhofer Institute for Algorithms and Scientific Computing University of Leeds Leibniz University of Hanover Flexiant British Telecom Arsys Total cost: 10.533K 2

OPTIMIS Scenarios 3

SLA use in OPTIMIS Provide SLA- aware infrastructure services with extended QoS capabilities SLAs between infrastructure providers and service providers Support for self- * and elasticity Definition and standardisation of new term languages Trust Risk Eco- efficiency Cost Data security & legal constraints SLAs based on Service Manifest of the Service Provider 4

OPTIMIS SLA Management Cloud QoS OPTIMIS SLA Management Standard compliant SLA Management component Full support of the OGF WS- Agreement specification SLA Negotiation supported through WS- Agreement Negotiation Based of the WSAG4J SLA Management Framework developed at Fraunhofer SCAI Functionality of the SLA Management Layer Mediates between SDO and IP components Negotiates IP services and triggers deployment Basic integration with IP components for service deployment implemented Integration of SLA monitoring planned for year 2 Innovations Support of TREC parameters as part of the SLA Enables self- management and self- optimization operations based on TREC parameters Dynamic SLA creation, monitoring and accounting Rule- based definition of SLA guarantees Dynamic detection of SLA compliance and violations Trigger self- management and self- optimization actions in order to prevent SLA violations Triggers accounting in case of SLA violation 5

SLA Management SLA Creation Service Deployment Optimizer GetTemplates Templates Infrastructure Provider SLA Management (Cloud QoS) Agreement Factory Facade Get Templates Lookup Agreement Factory Resource Proper5es Templat es Populate SLA Validation Template based creation of SLAs Templates contain restrictions on valid agreement offers Structural restrictions, i.e. which element must be contained in the offer and how often Create Agreement Create Agreement Agreement Offer Valida5on Validate Lookup Agreement Offer Validator Configura5on Value restrictions, i.e. what are valid values for specific elements Enforcement of offer integrity Constraint validation before service instantiation Protection of IP components Service Instan5a5on Instantiate Service Provisioning System Service Instantiation Admission of services done by IP admission control component Agreement EPR Agreement Crea5on Configure Deployment agreed services initiated by IP cloud optimizer

SLA Management Monitoring Agreement Properties Dynamic Resource Properties Static Resource Properties Terms Service Description Term 1... Service Property 1 Service Property 2 Guarantee Term Agreement State Service Term States Service Term State 1 State Guarantee Term States Guarantee Term State Agreement 2. update service terms states 4. update guarantee states compute states Agreement Monitor Guarantee Evaluator 3. evaluate guarantees 1. retrieve service states 5. notify Monitoring System Self- Management / Self Op5miza5on Accoun5ng SLA Monitoring Periodical monitoring of deployed services and guarantees Guarantee states are computed based on definition in SLA Guarantees can specify either an importance or monetary values Penalties and rewards with monetary values result in notifications to an accounting system Penalties and rewards with an importance result in notifications to the self- management and self- optimization components

SLA Management Monitoring of SLAs Agreement Static Properties Service Description... Specify the user requirements on the service provisioning process Dynamic Properties Static Properties Terms Service Description Service Property 1 Service Property 2 Guarantee Agreement State Service Term States Service Term State 1 State Guarantee Term States Guarantee Term State agreed service property actual service property Can be changed by the user with respect to the agreement creation constraints Service Properties Define variables in order to resolve static and dynamic property values Are used to define the service level objectives in SLA guarantees Guarantees Specify service level objectives Define self management events and accounting models in form of penalties and rewards Dynamic Properties Service Term States Dynamically generated during SLA monitoring Contain detailed information on the current service delivery Guarantee Term States Are automatically evaluated by the SLA management Accounting and Self- management capabilities are triggered based on evaluation results

SLA Negotiation 9

OPTIMIS Service Manifest <opt:elasticitysection> Describes the SP view on the Services <ws:servicedescriptionterm <opt:rule> ws:name="trec" ws:servicename="multipleimages"> <opt:scope> <opt:trecsection> <opt:trustsection> <opt:virtualsystemgroupid>instance-jboss-group</opt:virtualsystemgroupid> OVF XML template <opt:trustlevel>5</opt:trustlevel> </opt:scope> </opt:trustsection> Base for the Service <!-- Current <opt:risksection> amount of JVM Threads in the JBoss JVM --> Description Terms of the <opt:availabilityarray> <opt:kpiname>threadcount</opt:kpiname> <!-- defines a minimum availability of the VM of 98% Service at a day Level --> <!-- These values <opt:availability are not used for opt:assessmentinterval="p1d">98</opt:availability> this measurement --> Agreement between SP <opt:window>p5m</opt:window> <!-- defines a minimum availability of the VM of 99% at a month --> <opt:availability opt:assessmentinterval="p1m">99</opt:availability> and IP <opt:frequency>1</opt:frequency> </opt:availabilityarray> <!-- 100 users per thread, with a tolerance of 5% (so trigger Each at more section than in 105 the users, OVF </opt:risksection> and down again at below <opt:ecoefficiencysection> 95 users ) --> template is mapped to <opt:quota>100</opt:quota> <opt:leedcertification>notrequired</opt:leedcertification> one of the Service <opt:breeamcertification>notrequired</opt:breeamcertification> <opt:tolerance>5</opt:tolerance> Description Terms in the <opt:eucoccompliant>false</opt:eucoccompliant> </opt:rule> <opt:energystarrating>no</opt:energystarrating> Agreement template </opt:elasticitysection> </opt:ecoefficiencysection> <opt:costsection> <opt:dataprotectionsection> <opt:price opt:currency="eur">100.00</opt:price> <opt:dataprotectionlevel>dpa</opt:dataprotectionlevel> </opt:costsection> </opt:trecsection> </ws:servicedescriptionterm> <opt:dataencryptionlevel> <opt:encryptionalgoritm>aes</opt:encryptionalgoritm> </opt:dataencryptionlevel> </opt:dataprotectionsection> 10

OPTIMIS SLA Negotiation 11

Service Details Examples (TREC, Elasticity, Data Protection) Trust TrustLevel 5 Risk minimum availability of the VM of 99% at a month Eco Efficiency No certification required Cost 100.00 Euro (maximum cost) Elasticity (jboss instances) 100 users per thread, with a tolerance of 5% (so trigger at more than 105 users, and down again at below 95 users Data Protection Data Protection Level DPA Encryption Algoritm AES 12