Adaptive Computing Using PlateSpin Orchestrate

Adaptive Computing Using PlateSpin® Orchestrate

Gryphon McArthurSoftware Engineer [email protected]

© Novell, Inc. All rights reserved.2

Abstract

Adaptive Computing: Goes beyond just intelligently utilizing available resources; it

encompasses quality of service (QoS) targets, fault tolerance (high availability), monitoring, and iterative analysis of the resulting dataset to determine what corrective measures (adaptations) should occur at any given moment. As virtualization becomes widespread in the data center, the need for automating the placement and configuration of workloads (virtual machines) using an adaptive computing model becomes vitally important. This session demonstrates how to use events, introduced in PlateSpin® Orchestrate 2.0.2, to create rules that trigger workload provisioning, migration, and other virtual machine life-cycle operations.


Overview

Overview of PlateSpin® Orchestrate

Adaptive Computing and IWM

A Real-world Scenario

Steps to implement the solution

Advanced Workload Management

Thermal load balancing

Conclusion


Overview of PlateSpin® Orchestrate

Orchestration Server

VMHost systems running the Orchestration Agent

Graphical clients for:

Virtual machine management

Development of policies, jobs, etc.


Adaptive Computing

Definition of Adaptive Computing:

“Adaptive computing focuses on the methodology and

implementation of systems that adjust to different situations.

An adaptive system may change its own behavior to the

goals, tasks, interests, and other features of individual

users and the environment. Adaptivity is important for

ubiquitous and pervasive computing.”

Helsinki Institute for Information Technology


Intelligent Workload Management

Intelligent Workload Management enables IT organizations to manage and optimize computing resources in a policy-driven, secure and compliant manner across physical, virtual and cloud environments to deliver business services for end customers.

IntelligentWORKLOADManagement


Scenario/Goal

Automatically migrate VM's off of a VM host once the host's load exceeds 75% capacity for more than 5 minutes.

Constantly monitor/evaluate VM host loads

Migrate workloads to other hosts when load is too high

Allow workloads to return once load returns to normal


ImplementationCreate and Configure Objects using the “Development Client”


Implementation Steps

Step #1: Setup monitoring on VM hosts

Automatically configured when installed with PlateSpin® Orchestrate agent

Step #2: Create an RRD metric definition

Use the “load_one” metric with a 5 minute aggregation period

Step #3: Create an event

Step #4: Create the job for handling the event

Step #5: Create a schedule and trigger

for the event created in step #3 and the job created in step #4.


Deploying RRD Metrics DefinitionMetric Editor in the “Development Client” GUIDefining a “load_one” metric (RRD data aggregation)


Creating the EventEvent Editor in the “Development Client” GUIDefining policy/constraints for triggering a migration


Creating the JobJob Editor in the “Development Client” GUIUsing JDL to migrate VM's from a VM host


Connecting It All Together

Using the “Development Client” GUI

– Create the schedule

– Create the trigger

– Select the event

– Associate the job

Optionally:

– Put everything in a single “.job” file so it all deploys automatically


Creating the Schedule/TriggerScheduler View in the “Development Client” GUIAllows the “migrate” job to be invoked by the event


Example in ActionVM Hosts View in the “Development Client” GUIWorkloads are migrated away until load returns to “normal”


Advanced Workload Management

Thermal Load BalancingA “proof of concept” created by Adam Spiers

Addresses the problem of “hot spots” in the data center using a novel software-based solution.


The Problem: Data Center Cooling

• Cooling is expensive– In a typical data center, more power is spent on cooling than

on servers. Estimates range from 44%1 to as high as 63%.2

• Cooling is unavoidable– 10% of racks have ambient temperatures of 75°F or higher at

the air intake at the top of the rack. High temperatures are causes of decreased hardware reliability. Intermittent ghosts and outright hardware failures are three times more prevalent in the top third of racks than the bottom two-thirds.3

• Air conditioning units are critical to service availability– Servers can redline within 90 seconds of an AC unit failure.4

1 See, Simon. Is there a pathway to a Green Grid? Sun, 2008 http://www.ibergrid.eu/2008/presentations/Dia%2013/4.pdf2 http://h20331.www2.hp.com/ERC/cache/438048-0-0-225-121.html3 Sullivan, Robert F., Ph.D. Reducing Bypass Airflow Is Essential for Eliminating Hotspots. Upsite Technologies http://www.42u.com/data-center-hot-spots.htm4 Sharma, Ratnesh et al. Balance of Power: Dynamic Thermal Management for Internet Data Centers. HP Labs, 2003

http://www.hpl.hp.com/techreports/2003/HPL-2003-5.pdf

http://www.ibergrid.eu/2008/presentations/Dia%2013/4.pdf

http://h20331.www2.hp.com/ERC/cache/438048-0-0-225-121.html

http://www.42u.com/data-center-hot-spots.htm



The Problem: Data Center Cooling(Continued)

• Many data centers cool incorrectly

– Cooling over-capacity is very common, and is not a predictor of successful cooling. In one study, nineteen rooms studied ran on average 2.7 times more cooling equipment than required to cool the computer heat load. Two rooms ran 16 times more cooling than required, yet one had 20% hot racks/cabinets and the other had 7% hot racks/cabinets.1

• There are over 80 energy efficiency incentive or rebate programs offered by local utilities or state energy efficiency programs in the US alone.2

1 Sullivan, Robert F., Ph.D. Reducing Bypass Airflow Is Essential for Eliminating Hotspots. Upsite Technologies http://www.42u.com/data-center-hot-spots.htm2 The green data center. IBM, May 2007 http://www-900.ibm.com/cn/systems/migratetoibm/pdf/Energy-file03_OIW03002USEN.pdf

http://www.42u.com/data-center-hot-spots.htm

http://www-900.ibm.com/cn/systems/migratetoibm/pdf/Energy-file03_OIW03002USEN.pdf


A Novel (Novell®) Solution:Dynamic Thermal Load Balancing• Virtualization and live migration allows dynamic

relocation of workloads with no impact on service– Policy-based migration of VMs from hot spots to cool spots

• Energy consumption can be reduced by more than 14% by intelligent workload placement.1

• Risk of service outage due to computer room air conditioning unit failures can be mitigated by migrating workloads away from the failed unit.1

• Automatic VM live migration based on thermal policies can be implemented easily using PlateSpin® Orchestrate from Novell.

1 Sharma, Ratnesh et al. Balance of Power: Dynamic Thermal Management for Internet Data Centers. HP Labs, 2003 http://www.hpl.hp.com/techreports/2003/HPL-2003-5.pdf



Thermal Load Balancing

“Rack” of VM host hardware operating below capacity

Cooling Units

VM host hardware operating at or near capacity


VM MigrationVM Hosts View in the “Development Client” GUIWorkloads are migrated away from the “hotspots”


Conclusion

PlateSpin® Orchestrate provides the policy based data center automation capabilities needed to implement real-world adaptive computing scenarios

Orchestrate provides a set of tools and extensible framework for implementing your own unique solutions

This is just one component of Intelligent Workload Management

Questions and Answers

Unpublished Work of Novell, Inc. All Rights Reserved.This work is an unpublished work and contains confidential, proprietary, and trade secret information of Novell, Inc. Access to this work is restricted to Novell employees who have a need to know to perform tasks within the scope of their assignments. No part of this work may be practiced, performed, copied, distributed, revised, modified, translated, abridged, condensed, expanded, collected, or adapted without the prior written consent of Novell, Inc. Any use or exploitation of this work without authorization could subject the perpetrator to criminal and civil liability.

General DisclaimerThis document is not to be construed as a promise by any participating company to develop, deliver, or market a product. It is not a commitment to deliver any material, code, or functionality, and should not be relied upon in making purchasing decisions. Novell, Inc. makes no representations or warranties with respect to the contents of this document, and specifically disclaims any express or implied warranties of merchantability or fitness for any particular purpose. The development, release, and timing of features or functionality described for Novell products remains at the sole discretion of Novell. Further, Novell, Inc. reserves the right to revise this document and to make changes to its content, at any time, without obligation to notify any person or entity of such revisions or changes. All Novell marks referenced in this presentation are trademarks or registered trademarks of Novell, Inc. in the United States and other countries. All third-party trademarks are the property of their respective owners.

Documents

Adaptive Computing Using PlateSpin Orchestrate