Forecast 2014: Maximizing Your Cloud Service Quality and Costs

MAXIMIZING YOUR CLOUD

SERVICE QUALITY AND COSTS

1

Ed Simmons- Executive Director, Platform Engineering, UBS - ODCA Infrastructure Workgroup Co-Chair

Ray Solnik- President, Appnomic Systems, Inc.- ODCA Infrastructure Workgroup Member

OBJECTIVE AND AGENDA

Objectives: 1. For you to understand how Proofs of Concept Work at the ODCA and why

you should be interested in participating2. For you to learn and benefit from one of the POCs implemented this year

on behalf of ODCA

Agenda:1. The ODCA Usage Models and the POC2. POC Overview3. POC Real Life Results4. Q&A

2

Finer Grain Usage Models

Master Usage ModelsPUBLISHED USAGE MODELS

• Virtual Machine Interoperability• Long Distance Migration• IO Control• Service Catalog• Standard Units of Measure • Provider Assurance• Security Monitoring• Regulatory Framework• Security Data Framework• Data Security• Data Management

• Carbon Footprint• SaaS Interoperability• PaaS Interoperability• Interoperability across Clouds• Identity Management Interoperability• Cloud Based Identity Provisioning• Cloud Based Identity Governance and Auditing• Single Sign on Authentication• IaaS Privileged User Access• SW Entitlement Management Framework

Usage Model Contents:

• Exec Summary

• Problem Statement

• Use Cases

• RFP Questions

• Industry Action

Compute IaaS 1.0, Scale-out Storage 1.0, Software-Defined Networking 1.0

Service Orchestration 1.0

Commercial Framework 1.0

Information as a Service 1.0

PROOF OF CONCEPTSODCA Proof of Concepts (POC) Objectives:1. Validate & improve UMs through real world testing2. Lead to implementation: trigger real adoption of UM based solutions3. Influence product roadmaps: promote innovation in line with ODCA

requirements4. Showcase provider solutions & services5. Increase visibility for the alliance & participants through press & publications

4

PoC Type Description

SP/UM Alignment Paper evaluation or lightweight review performed in lab/online, generate Cloud Finder data, publish paper

Biz Case Study Set up usage scenario(s), execute tests, collect and analyze data, provide recommendations, publish raw internal report and public best practices paper

POC PROCESS

ADOPTION

• Identify UM• Define PoC scope &

objectives • Secure WG alignment &

sponsor• Understand governance

• Create project plan• Identify team roles &

responsibilities, tasks

• Scope Resource Requirements (people, technology)

• Create PoC Team

CONCEPT RESOURCE ALIGNMENT PROJECT PLAN

Planning1 2 3

• Deliver project work• Document results (raw

results/report)

• Develop public collaterals• Socialize with stakeholders• Publish through ODCA

Marketing with PR activities

PROJECT EXECUTION RESULT COLLATERALS

Execution 4 5

5

POC CASE STUDY: COMPUTE IAAS / SUOM

Measured:• Actual end user transaction volumes• Transaction response times• CPU capacity utilization• Memory utilization

Learned:

• Explored methods for comparing multiple cloud IaaS providers

• Insights can be used to tune workloads and costs for particular cloud providers

• New operations & troubleshooting techniques

• Industry gap in comparison shopping -recommending “Cloud Facts” labelling

Value of Participating:

• Disney• First-hand experience and

learnings from new ITOA solution• Influence maturity of cloud using

real-world experience

• Appnomic• Accelerate solution & adoption• Build awareness to generate sales

6

CASE STUDY: HOW DO I GET AND ENSURE ENTERPRISE GRADE CLOUD PERFORMANCE AT THE BEST PRICE?

Participants: Appnomic, Disney, and Intel

Goal: Establish a clear, equitable means of rating cloud service providers in respect to performance and quality of service

Business Value: Provide method for enterprise decision-makers to evaluate competing services, including their own production application environments, achieve the best price/performance available

Approach: Run synthetic transactions through a Disney app in an enterprise private cloud and two external providers; profile environments using Appnomic AppsOne IT Operations Analytics platform; “rinse repeat.”

Full report available at: http://www.appnomic.com/ODCA_ITOA_POC

7

THE COMPETITORS: EQUIVALENT COST

8

$1.06/hr $1.12/hr

THE RESULTS: NOT EQUIVALENT PERFORMANCE

9

Transaction volumes completed at 3 different tiers of load

Hour of the Day Hour of the Day Hour of the Day

Tran

sact

ion

Cou

nt

CIaaS Platform 1Completed Transactions

CIaaS Platform 2Completed Transactions

CIaaS Platform 3Completed Transactions

   Metric 

CIaaS Platform 1

CIaaS Platform 2

CIaaS Platform 3  Comments 

End‐User 

Metrics  Peak Transaction Volume (TPM)  75,000  47,000  25,000 

The difference between 75,000 and the values in this table are transaction requests that were not completed.

Peak Load Transaction Response Time (milliseconds) 

750  1,200  720 

CIaa

S Mea

sures 

Peak App Server CPU Utilization   2.4%  13.0%  8.5% 

Peak App Server Memory Utilization  26.0%  99.0%  11.5% While CIaaS Platform 1 is higher (worse than) CIaaS Platform 3, it remains in an acceptable range. 

Cost Per Hour See Comment 

$1.06  $1.12 CIaaS Platform 1 is an internal infrastructure and does not have a per‐user cost structure associated with it. 

DASHBOARD SUMMARY

PRODUCTION APPLICATION USAGE PATTERNS (AUPS) CAN ACT AS THE BASELINE CONTROL

FIGURING OUT WHAT TO PROVISION FROM WHICH PROVIDER IS CURRENTLY VERY CHALLENGING

ODCA has evaluated other industry solutions and is investigating the equivalent of the food industry solution with helpful labelling

THE “CLOUD FACTS” LABEL

THANK YOU

Ed Simmons, Executive Director, Platform Engineering, [email protected]

Ray Solnik, President, Appnomic Systems, [email protected]

ADVANCED IT ANALYTICS DEFINEDFor very complex IT environments, the automated or computing-machine organization and interpretation of massive volumes of metrics representing the IT environment into meaningful, actionable insights or automated actions to improve IT systems’ performance and/or security.Often, outcomes can have material positive impact on business results through prevention of IT operations incidents like slowness and outages as well as enabling many other use cases based on the various mathematical algorithms and techniques that may be utilized.

h t t p : / / d o u g m c c l u r e . n e t / b l o g / d e f i n i n g - i t - o p e r a t i o n s - a n a l y t i c s /

M u l t i p l e t e r m s : I T O p e r a t i o n s A n a l y t i c s ( G a r t n e r ) , I T A n a l y t i c s ( F o r r e s t e r ) , A d v a n c e d A n a l y t i c s ( E M A ) . N o t A P M – A P M i s a d o p t i n g I T O A .

16

© 2 0 1 4 O p e n D a t a C e n t e r A l l i a n c e , I n c . A L L R I G H T S R E S E R V E D .