Enhancing Energy Efficiency in Cloud Management through Code Annotations and the Green Abstraction...

21
Enhancing Energy- Efficient Cloud Management ..through Code Annotations and the Green Abstraction Layer RAFFAELE BOLLA MATTEO REPETTO LUIGI SAMBOLINO DANILO TIGANO UCC 2015 – 7° DECEMBER 2015 – LIMASSOL, CYPRUS

Transcript of Enhancing Energy Efficiency in Cloud Management through Code Annotations and the Green Abstraction...

Page 1: Enhancing Energy Efficiency in Cloud Management through Code Annotations and the Green Abstraction Layer

Enhancing Energy-Efficient Cloud

Management..through Code Annotations and the Green

Abstraction Layer

RAFFAELE BOLLA MATTEO REPETTO LUIGI SAMBOLINO DANILO TIGANO

UCC 2015 – 7° DECEMBER 2015 – LIMASSOL, CYPRUS

Page 2: Enhancing Energy Efficiency in Cloud Management through Code Annotations and the Green Abstraction Layer

Outline

1. Motivation

2. Arcadia’s Vision

3. Our approach

4. Ongoing work

5. Conclusions

Page 3: Enhancing Energy Efficiency in Cloud Management through Code Annotations and the Green Abstraction Layer
Page 4: Enhancing Energy Efficiency in Cloud Management through Code Annotations and the Green Abstraction Layer

Power usage effectiveness

Google Facebook Amazon Microsoft Yahoo Ebay0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

https://gigaom.com/2012/03/26/whose-data-centers-are-more-efficient-facebooks-or-googles/

Author
Power usage effectiveness (PUE) is a measure of how efficiently a computer data center uses energy; specifically, how much energy is used by the computing equipment (in contrast to cooling and other overhead).PUE is the ratio of total amount of energy used by a computer data center facility to the energy delivered to computing equipment.PUE was developed by a consortium called The Green Grid. PUE is the inverse of data center infrastructure efficiency (DCIE). An ideal PUE is 1.0. Anything that isn't considered a computing device in a data center (i.e. lighting, cooling, etc.) falls into the category of facility energy consumption. \mathrm{PUE} = {\mbox{Total Facility Energy} \over \mbox{IT Equipment Energy}}
Page 5: Enhancing Energy Efficiency in Cloud Management through Code Annotations and the Green Abstraction Layer

The consolidation problem

1234

ChallengesVMs impredictabilityWorkload variability

SolutionsScale frequenciesScale powered componentsReallocation of VMs

Author
Each Virtual Machine represent a Service
Page 6: Enhancing Energy Efficiency in Cloud Management through Code Annotations and the Green Abstraction Layer

Reallocation TimeDowntimeSLAQoSQoE

SolutionExploit applications behavior knowledge in data centers

Efficiency Trade-Off

Author
Page 7: Enhancing Energy Efficiency in Cloud Management through Code Annotations and the Green Abstraction Layer

DevOps philosophy

• Communication• Collaboration• Process

automation• Infrastructure

changes

Author
DevOps philosophy tries to set a tighter collaboration between software developers and operation staff. Developers are asked to include information to let the applications be aware of their context, or automatic configuration.
Page 8: Enhancing Energy Efficiency in Cloud Management through Code Annotations and the Green Abstraction Layer

European Project

Helping services placement in the cloud

For Developers, DevOps and Providers

Page 9: Enhancing Energy Efficiency in Cloud Management through Code Annotations and the Green Abstraction Layer

ARCADIA Conceptual Architecture

Author
Meta Model dovrebbe essere la struttura delle applicazioni, fatta di di "contenitori" base e "incastri" per poter creare delle cateneThe meta-model template is a collection of properties, relationships, attributes, and policies related to the specific app or service chain. It's written by Developer directly as source-code level and assisted and validated by Arcadia IDE-Plugin. More details are going to be given in next slide talking about TOSCA.
Author
Il Context Model e` quello che dovrebbe sviluppare ARCADIA, ovvero una astrazione del contesto di esecuzione: CPU, memoria, storage, network, energy efficiency, QoS, Sistemi Operativi, dipendenze software ed hardware, attuale carico di lavoro, parametri da monitorare, e non so quanto altro possa essere utile specificare per la corretta esecuzione dell'applicazione.
Page 10: Enhancing Energy Efficiency in Cloud Management through Code Annotations and the Green Abstraction Layer

Our idea

1. Meta-models exploitation

2. Policies

3. Unused replicas clustering

4. Sleeping Nodes

Page 11: Enhancing Energy Efficiency in Cloud Management through Code Annotations and the Green Abstraction Layer

Smart Controller

Page 12: Enhancing Energy Efficiency in Cloud Management through Code Annotations and the Green Abstraction Layer

Labelled Grouping

Light

Dynamic

Critical

Page 13: Enhancing Energy Efficiency in Cloud Management through Code Annotations and the Green Abstraction Layer

Ongoing Work

1. Monitoring

2. Profiling

3. Consolidation Algorithms

Page 14: Enhancing Energy Efficiency in Cloud Management through Code Annotations and the Green Abstraction Layer

Placement AlgorithmPower consumption minimization

Sets of elements considered: Physical Machines Virtual Machines Network Machines Set of cooling machines Links Traffic demand

Page 15: Enhancing Energy Efficiency in Cloud Management through Code Annotations and the Green Abstraction Layer

Algorithm Constraints

1. Maximum number of re-allocations

2. Link capacity depends on interfaces power states

3. Maximum number of hops between VMs

Page 16: Enhancing Energy Efficiency in Cloud Management through Code Annotations and the Green Abstraction Layer

TestBed

Page 17: Enhancing Energy Efficiency in Cloud Management through Code Annotations and the Green Abstraction Layer

The Green Abstraction Layer1. ETSI standard interface (ES 203 237)

2. Developed by ECONET Consortium

3. Power management in networking devices

4. GALv2

5. Basic Primitives

6. Interaction between heterogeneous HW

7. Power States

Page 18: Enhancing Energy Efficiency in Cloud Management through Code Annotations and the Green Abstraction Layer

GAL

Page 19: Enhancing Energy Efficiency in Cloud Management through Code Annotations and the Green Abstraction Layer

GAL Commands

Discovery Provisioning Release Monitoring Commit Rollback

Page 20: Enhancing Energy Efficiency in Cloud Management through Code Annotations and the Green Abstraction Layer

Questions?

Page 21: Enhancing Energy Efficiency in Cloud Management through Code Annotations and the Green Abstraction Layer

Thank you