65pt Intel Clear PRO Presentation Title

advertisement

Kingsum Chow, Principal Engineer, Intel

Sunil Raghavan, Senior Software Engineer, Intel

Saurabh Dixit , Director, Oracle

Legal Disclaimer

Intel technologies’ features and benefits depend on system configuration and may require enabled hardware, software or service activation. Performance varies depending on system configuration. No computer system can be absolutely secure. Check with your system manufacturer or retailer or learn more at intel.com.

Software and workloads used in performance tests may have been optimized for performance only on Intel microprocessor Performance tests, such as SYSmark and MobileMark, are measured using specific computer systems, components, software, operations and functions. Any change to any of those factors may cause the results to vary. You should consult other information and performance tests to assist you in fully evaluating your contemplated purchases, including the performance of that product when combined with other products. For more complete information about performance and benchmark results, visit www.intel.com/benchmarks .

Tests document performance of components on a particular test, in specific systems. Differences in hardware, software, or configuration will affect actual performance. Consult other sources of information to evaluate performance as you consider your purchase.

For more complete information about performance and benchmark results, visit http://www.intel.com/benchmarks

Intel, the Intel logo, Intel Inside, Xeon are trademarks of Intel Corporation in the U.S. and/or other countries. *Other names and brands may be claimed as the property of others.

© 2015 Intel Corporation.

2

3

Keep Learning with Oracle University

Classroom Training

Learning

Subscription

Live Virtual Class

Training On

Demand

Cloud

Technology

Applications

Industries

education.oracle.com

4

Session Surveys

Help us help you!!

We invite you to take a moment to give us your session feedback. Your feedback will help us to improve your conference.

Please be sure to add your feedback for your attended sessions by using the

Mobile Survey or in Schedule Builder.

55

Agenda

Oracle Public Cloud

• Overview and Architecture

• IaaS, PaaS, Storage & Networking

Optimizing the Cloud

• Project Apollo

• Intel® Xeon® Processor E5-2600 v3 Product Family

• Early Results

A spoonful of analytics

• Analytics for Cloud performance optimization

6

Oracle Compute Cloud Service Delivers

Overview and Architecture

Core OCCS

OCCS is Foundation for New Oracle

PaaS/SaaS Services

77

Oracle Infrastructure-as-a-Service

Secure. Reliable. Low Cost.

Storage

Elastic Storage

Compute

Dedicated Compute

Network

Software-defined Network

IaaS: General Purpose, Engineered Systems

88

Compute: Detailed Features

Compute Lifecycle Management Secure Access

Hardware Isolation

Launch Dedicated Instances on single-tenant hardware with network isolation

Active VM Recovery

Configure HA Policies to automatically recover failed

VMs

Instance anti-affinity

Control instance placement and distribute workloads

OCPU Allocation

OCPUs not over subscribed providing predictable performance

Image Management

Provision virtual machines using pre-packaged images or build your own images

Elastic Block Storage

Store data and applications in persistent block volumes.

Maintain persistence at OS-level using Bootable storage volumes

Orchestration Management

Automate provisioning and lifecycle operations of virtual compute topologies

Metering & Quota

Management

Monitor usage and performance metrics

Secure SSH Access

Access Oracle Compute

Cloud instances from a remote host by using a secure shell

Dynamic Firewall

Control network traffic among individual instances and/or between groups of instances

Federated Identity

Integrate with MS AD/LDAP so users can login through single sign-on

Virtual Networking

Fast Connect

Bypass public internet over a private connection. Co-locate workloads and cross connect with partner exchange.

Site to Site VPN

Connect on-premise resources to dedicated compute zones in Oracle

Public Cloud

Reserved IP

Connect each instance to the internet and access Oracle

Public Cloud resources from anywhere

99

Oracle Platform-as-a-Service

Database Java Developer Mobile Documents

Social

Network

Big Data

PaaS Service Manager

BI

IaaS API

Messaging Process Integration

Block Storage Object Storage Compute Network

10

PaaS Services Options

Customer managed services

 Customer has VM access & full control over the VM,

 Can customize settings and controls patching and upgrade cycle

Oracle managed services

 Services are fully managed by Oracle

 Oracle ensures all necessary upgrades are applied

 Both options are already tuned out of box for best performance

 Self-managed, with fully automated cloud tooling for admin, LCM operations

Oracle Storage Cloud Services

RMAN

Storage

Gateway

OpenStack SWIFT API

Eventual Consistency

Any NAS or SAN

 Accessible

 Secure : Enterprise-grade data protection and privacy policies

 Scalable : ON-Demand Capacity

 Reliable : Redundancy policies to Data HA

 Standards-Based : OpenStack Swift compatible REST API & Java library for data management

 Hybrid storage tiers

Archive & Glacial QOS • Backup

Global Namespace  User backup, PaaS & IaaS backup target

• Archive

 Archive for long-term retention , compliance needs

Virtual Networking: IP Reservations and Pools

Connecting to the Internet

IP Reservation

Persistent Public IP

Reservation for VM instances

Elastic IP

Ephemeral Public IP from shared IP Pool

IP Pool

Pool of Public IP Addresses {

192.0.2.0/24, 198.51.100.0/24,

203.0.113.0/24

}

Virtual Networking: Security Rules

Security Rule = Security List or IP List + Applications + Security List

Source list

Destination list

Security List

A group of VMs .

Security Application

Abstracts the details of protocol definitions and port number ranges. i.e. [ssh, port 22, tcp/udp]

OR Security list

A group of VMs.

Security IP List

A group of subnet /

IP addresses.

Defines how a source security or security IP list can connect to a destination security list via a specific security application or protocol. Only PERMIT rules allowed.

Project Apollo

Project Apollo

- Deliver predictable high performance for applications running on Oracle

Cloud

- Characterize the cloud using Oracle

Cloud workloads

- Optimize the cloud to deliver maximum performance for the workload

- Innovate, develop new technologies

- Generate blueprint of an optimized data center

Components in Cloud

Application

Developer

Java

Application

Server

Administrator

VMs

Virtual

Storage

Network

Load

Balancer

Security

Service Provider

Servers Storage Switches Racks

Database Platform

Monitoring Infrastructure

Power and

Thermal

Hardware

17

What does Cloud performance mean?

Application

Developer

Application performance – Compute,

Storage and Network,

Scalability…

Administrator

Multi-tenancy, Predictable performance,

Security, Elasticity,

Composability, High Availability …

Service Provider

Optimal resource usage,

Power, Space, Cooling …

18

Cloud setup at Intel

Infrastructure, Power and Cooling

Cloud Incubator at Intel

2000 compute cores

27 TB of RAM

100 TB of Storage

Java as a Service

Intel Xeon E5 – 2699 v3

Database as a Service

Oracle ZFS Storage

Appliance

Operating system /

Hypervisor

Oracle Cloud

Platform

19

Extend to other Intel Cloud Technologies

Future plans to evaluate Intel Cloud technologies in Memory, Storage,

Network, Security, Power….

Reference1: http://newsroom.intel.com/community/intel_newsroom/blog/2015/07/28/intel-and-micron-produce-breakthrough-memory-technology

Reference2: http://www.intel.com/content/www/us/en/architecture-and-technology/intel-rack-scale-architecture.html

Intel® Xeon® Processor E5-2600 v3 Product Family

Intel ® Xeon ® Processor E5-2600 v3

• New Haswell microarchitecture with Intel® AVX 2.0

• Up to 50% more cores and cache for up to 50% performance increase 1

• New capabilities in the areas of virtualization, security and power efficiency

• DDR4 memory support to scale performance

Feature

Cores/Threads per socket

Last-level Cache (LLC)

Max Memory Speed (MT/s)

QPI Speed (GT/s)

Max DIMM Capacity

PCIe* Lanes /

Controllers/Speed

TDP (W)

AVX

Xeon ® processor E5-2600 v2 product family

Up to 12 Cores / 24 Threads

Up to 30 MB

Xeon ® processor E5-2600 v3 product family

Up to

Up to 1866

2x QPI 1.1 channels 6.4, 7.2, 8.0

Up to 12 Slots/Processor

130, 115, 95W

AVX 1.0

18 Cores /

Up to 45 MB

Up to 2133

2x QPI 1.1 channels 6.4, 8.0, 9.6

Up to 40 / 10 / PCIe* 3.0 (2.5, 5, 8 GT/s)

36 Threads

135, 120, 105W

AVX 1.0, AVX 2.0

Software and workloads used in performance tests may have been optimized for performance only on Intel microprocessors. Performance tests, such as SYSmark and MobileMark, are measured using specific computer systems, components, software, operations and functions. Any change to any of those factors may cause the results to vary. You should consult other information and performance tests to assist you in fully evaluating your contemplated purchases, including the performance of that product when combined with other products. NOTE: 1 See benchmark published results on this slide and configurations on slide following. For more information go to http://www.intel.com/performance/datacenter

Intel ® Xeon ® Processor E5-2600 v3 Product Family

IMPROVED

NEW

50% MORE

36% LESS

Faster

Memory

Virtualization features

Cores &

Threads

Idle power

Integrated IO (PCIe

3.0)

Security features

Last-level cache

Source: http://www.intel.com/content/dam/www/public/us/en/documents/product-briefs/xeon-e5-brief.pdf

Software and workloads used in performance tests may have been optimized for performance only on Intel microprocessors. Performance tests, such as SYSmark and MobileMark, are information and performance tests to assist you in fully evaluating your contemplated purchases, including the performance of that product when combined with other products. NOTE: 1 See benchmark published results on this slide and configurations on slide following. For more information go to http://www.intel.com/performance/datacenter

Early Results

Characterizing the Cloud

• Model Oracle Cloud workloads with multiple simultaneous applications

• Composed of:

• JaaS & DBaaS application workload

• CPU , IO & Network stress

• Real-time data gathered from across the stack

• Application performance

• Software logs – Java, Application Server, Database

• Cloud Platform / OS

• Statistics from VMs, Hypervisor

• Intel Performance Counters

• Storage appliance

• Switches

25

Early optimization results

We can achieve significant performance gains from our early optimization efforts of Oracle Cloud for Intel Xeon 2699 v3.

 Up to 1.5x for response time sensitive apps*

 Up to 1.2x

for throughput apps*

 Up to 10x better predictability*

– Less than 3% variation against up to 29% before optimization

1,6

1,4

1,2

1

0,8

0,6

0,4

0,2

0

Latency sensitive workload

Before optimization

Throughput sensitive workload

After optimization

* Software and workloads used in performance tests may have been optimized for performance only on Intel microprocessors. Performance tests, such as SYSmark and

MobileMark, are measured using specific computer systems, components, software, operations and functions. Any change to any of those factors may cause the results to vary. You should consult other information and performance tests to assist you in fully evaluating your contemplated purchases, including the performance of that product when combined with other products.

* For more information go to http://www.intel.com/performance/datacenter .

Better predictability and scaling

- By enhancing resource allocation mechanism we can achieve:

- More predictable performance*

- Linear scaling* - Linear increase in performance with increase in

OCPUs

700

600

500

400

300

200

100

0

0 5 10

OCPUs

15 20

* Software and workloads used in performance tests may have been optimized for performance only on Intel microprocessors. Performance tests, such as SYSmark and

MobileMark, are measured using specific computer systems, components, software, operations and functions. Any change to any of those factors may cause the results to vary. You should consult other information and performance tests to assist you in fully evaluating your contemplated purchases, including the performance of that product when combined with other products.

* For more information go to http://www.intel.com/performance/datacenter .

Use of analytics for cloud performance optimization

A Scenario for Analytics

Apps

App 1

App 2

Mid-Tier

Common

Common 1

 WLS Application Servers in the Cloud

 Simulated Users

 Data collected from Simulated Users and Servers

29

Approach

Raw data

Raw data

Python

Multiple Platforms for Processing

R

Tidy Data Format

Tidy Data Format

Merge Data sets over Time

Check Processing

Quality

Compute Statistics

Posterior Analysis

System Monitoring

Workload

Application

System/OS/VM

Microprocessor

• OATS

*

(e.g., no of users, response times)

• e.g. Java logs

• system activity report (sar)

• EMON

**

* OATS – Oracle Application Testing Suite. Testing solution for web applications using Oracle Applications and Oracle Database.

** EMON – Intel internal tool for logging event counters against a time base

Garbage Collection(GC) Data

2015-07-24T13:53:13.141-0700: 75.604: [GC [PSYoungGen:

1133359K->165347K(1223680K)] 1133447K->165470K(4020224K),

0.1085510 secs] [Times: user=0.59 sys=0.08, real=0.11 secs]

2015-07-24T13:53:22.445-0700: 84.909: [GC [PSYoungGen:

1214435K->168469K(1223680K)] 1214558K->168672K(4020224K),

0.1442510 secs] [Times: user=0.97 sys=0.14, real=0.14 secs]

2015-07-24T13:53:31.495-0700: 93.959: [GC [PSYoungGen:

1217557K->149712K(1199104K)] 1217760K->149923K(3995648K),

0.1272560 secs] [Times: user=0.75 sys=0.01, real=0.13 secs]

2015-07-24T13:53:35.700-0700: 98.163: [GC [PSYoungGen:

1198800K->145280K(1185792K)] 1199011K->145499K(3982336K),

0.0946850 secs] [Times: user=0.78 sys=0.02, real=0.10 secs]

2015-07-24T13:53:41.997-0700: 104.460: [GC [PSYoungGen:

1131904K->88361K(1192448K)] 1132123K->146072K(3988992K),

0.1296750 secs] [Times: user=1.03 sys=0.14, real=0.13 secs]

2015-07-24T13:53:51.739-0700: 114.203: [GC [PSYoungGen:

1074985K->118373K(1202176K)] 1132696K->228993K(3998720K),

0.2367950 secs] [Times: user=1.00 sys=0.09, real=0.24 secs]

2015-07-24T13:53:59.035-0700: 121.498: [GC [PSYoungGen:

1116261K->145330K(1193984K)] 1226881K->266899K(3990528K),

0.2270100 secs] [Times: user=0.59 sys=0.02, real=0.23 secs]

2015-07-24T13:54:03.826-0700: 126.289: [GC [PSYoungGen:

1143218K->53006K(1190912K)] 1264787K->233618K(3987456K),

0.0936990 secs] [Times: user=0.56 sys=0.09, real=0.10 secs]

Every application server has its own GC Log,

Many files in the cloud

Messy, semi-structured data ?

• Column headers are values, not variable names.

• Multiple variables are stored in one column.

• Variables are stored in both rows and columns.

• Multiple types of observational units are stored in the same table.

• A single observational unit is stored in multiple tables.

References http://vita.had.co.nz/papers/tidy-data.pdf

Data Processing, by “Hadley Wickham”

Chief Data Scientist at RStudio

Data Processing is the most essential part of data analysis. It encompasses activities like outlier detection, data parsing, missing value imputation, etc.

Hadley’s contribution to the Data Analytics society:

 Proposed a guideline for processed data-> “tidy data format”.

 Developed packages in R, that would make data processing easier.

dplyr, tidyr, ggplot

GC Logs + SAR + EMON + …

2015-07-16T10:27:20.150-0700 55.874 PSYoungGen1572864 81116 1835008 1572864 81196 3932160 0.07887

2015-07-16T10:27:32.854-0700 68.578 PSYoungGen1653980 131313 1835008 1654060 131409 3932160 0.132623

2015-07-16T10:27:49.518-0700 85.242 PSYoungGen1704177 220893 1835008 1704273 220997 3932160 0.222767

2015-07-16T10:27:59.456-0700 95.18 PSYoungGen1793757 159853 1835008 1793861 159965 3932160 0.157783

2015-07-16T10:28:08.902-0700 104.626 PSYoungGen1732717 169137 1835008 1732829 169321 3932160 0.129966

2015-07-16T10:28:20.469-0700 116.193 PSYoungGen1742001 230370 1719296 1742185 305359 3816448 0.341499

2015-07-16T10:28:26.412-0700 122.135 PSYoungGen1719266 160491 1793024 1794255 320956 3890176 0.189678

2015-07-16T10:28:31.324-0700 127.048 PSYoungGen1649387 162008 1818624 1809852 356704 3915776 0.117959

2015-07-16T10:28:38.325-0700 134.049 PSYoungGen1690840 51971 1807360 1885536 350114 3904512 0.162245

2015-07-16T10:28:49.314-0700 145.038 PSYoungGen1580803 113382 1812992 1878946 432032 3910144 0.101396

2015-07-16T10:28:56.206-0700 151.929 PSYoungGen1637094

2015-07-16T10:29:08.559-0700 164.283 PSYoungGen1566486

42774

44481

1566720

1802752

1955744

1982849

459137

493787

3663872 0.121342

3899904 0.09115

2015-07-16T10:29:53.981-0700 209.705 PSYoungGen1556929

2015-07-16T10:47:12.381-07001248.105 PSYoungGen1579052

2015-07-16T11:09:06.062-07002561.785 PSYoungGen1544331

2015-07-16T11:09:11.093-07002566.817 PSYoungGen1563726

2015-07-16T11:09:15.712-07002571.436 PSYoungGen1561650

66604

25739

45134

29234

37478

1804800

1806848

1807872

1816576

1812992

2006235

2058499

2061547

2091549

2131177

546051

542955

572957

598761

616113

3901952 0.120864

3904000 0.093274

3905024 0.052096

3913728 0.081865

3910144 0.050888

2015-07-16T11:09:17.518-07002573.242 PSYoungGen1569894 13784 1827328 2148529 611324 3924480 0.043828

2015-07-16T11:09:21.669-07002577.393 PSYoungGen1567192 47785 1823232 2164732 654786 3920384 0.059617

Tidy Data Format !!!

0.71

0.39

0.52

0.32

0.35

0.29

0.2

0.34

0.16

0.25

0.43

0.63

0.61

0.66

0.95

0.92

0.72

0.15

0.2

0.16

0.1

0.12

0.09

0.12

0.09

0.05

0.08

0.05

0.08

0.13

0.23

0.16

0.13

0.34

0.19

0.12

0.04

0.05

0.09

0

0.1

0.03

0.03

0.04

0

0.03

0

0.03

0.1

0.04

0.01

0

0.09

0.06

0.04

0.01

0

Advantages of Aggregating Tidy Data

Analysis made possible.

Data visualization becomes handy.

Easy correlation among the various metrics on different systems.

Comparison of trends of metrics across systems.

Summary

Oracle Cloud

Project Apollo: Joint Intel-Oracle Collaboration

Cloud Performance Analysis for Your Applications

• Data Collection

• Data Cleansing

• Analytics

Demo: INTERACTIVE ANALYTICS FOR CLOUD

PERFORMANCE

I will conduct the demo and hope to answer more questions tomorrow.

Kingsum Chow

2:30pm-4:00pm (Tuesday Oct 27)

Big Data Showcase, Moscone South, Booth 2223.

38

Download