C5 journey to_the_cloud_with_oracle_sparc

Copyright © 2015, Oracle and/or its affiliates. All rights reserved. |
Journey to the Cloud with Oracle SPARC
Modern Infrastructure for Legacy Unix Applications and the Cloud
Benny K. Chan
Principal Sales Consultant
Oracle Hong Kong, Systems
October 26, 2016

Copyright © 2016 Oracle and/or its affiliates. All rights reserved. |Copyright © 2016 Oracle and/or its affiliates. All rights reserved. |
Cloud Impacts Every Industry and Every Geography
2Copyright © 2016, Oracle and/or its affiliates. All rights reserved. |

Copyright © 2016 Oracle and/or its affiliates. All rights reserved. | 3
The Business Demands
Cloud for Speed and Agility
• Compelling applications
• Better workflow
• Deeper analytics
• Reduced costs

Few Options for
Modernizing Legacy
Business Applications
Choices are not good:
• Do nothing
• Change everything
What If There Were a Better Way?
4

Copyright © 2016 Oracle and/or its affiliates. All rights reserved. |
Deploy Legacy Apps on
Modern Cloud-Ready
Infrastructure
• Leverage investments
without starting over
• Improved security, capacity
• Flexibility, app development
• Simplified management and
service delivery
• Clear Path to the Cloud

Copyright © 2016 Oracle and/or its affiliates. All rights reserved. |Copyright © 2016, Oracle and/or its affiliates. All rights reserved. |
Leaders in Enterprise IaaS
6

Why Oracle for Business Critical Infrastructure
7
Only Oracle offers you a pathway to the cloud for business-critical applications
Modern
Platform
Cloud
Ready

Same System, On-premises or in the Oracle Cloud
8
Enterprise Cloud Infrastructure
Optimized Solution
SPARC Model 300
Dedicated compute
• Test/Dev
• Production
• Disaster Recovery
• Bursting
SPARC Systems for
Enterprise-grade Infrastructure

Copyright © 2016 Oracle and/or its affiliates. All rights reserved. | 910/26/2016
On-Premises /
Private Cloud
on
SPARC M7
Public Cloud
on
SPARC M7
Same System, On-premises or in the Oracle Cloud

Analytics
Unheard of Results for Apps & Cloud, Attainable Now
10
Security. Efficiency. Simplicity.

Security

A Couple of Famous Examples: Heartbleed and Venom
12
Silicon Secured Memory Protection From Read and Write Attacks
Buffer Over-Read Attack Buffer Over-Write Attack

More Secure and Reliable SW With Silicon Secured Memory
Memory
Other Processors
Pointer
No protection from stale
pointer and off end
access, buffer over-read,
buffer over-write
Memory
Pointer
SPARC M7 Processor
H/W compares pointer
“key” with memory “lock”
13

SPARC: Broadest Set Of Ciphers For All Your Apps
14
On-chip cryptography accelerators off-load cores: near zero performance cost
M7
32 Cores and
32 Crypto Accelerators
per Processor Clear Data In
Encrypted Data Out

End to End Security For Database, Apps, and Web Tiers
10,000 15,000
Only 2%
Impact
with
SPARC S7
UNSECURE
SECURE
5,000
14,121.46
EjOps
14,400.78
EjOps
15
SPECjEnterprise2010 (see disclosure slide)

Efficiency
Performance

SPARC M7: 32-Core Revolution
More Throughput Performance
More Single Thread Performance
More Encryption Bandwidth
More Cores – Higher Core Frequency - More Cache – More Cache per Core – More Memory
Bandwidth More IO Bandwidth - More PCIe Lanes – Lower Cache Latency – Lower Memory Latency
17
Data Analytics Accelerators
Silicon Secured Memory

Data Analytics Accelerator (DAX)
18
• 32 In-silicon offload engines on
M7
• Cores/Threads operate
synchronous or asynchronous
to offload engines
• User level synchronization
through shared memory
• High performance at
low power
• Up to 3x more memory
bandwidth than x86
Decompress
Unpack/
Alignment
Scan, Filter,
Join
Result
Format/
Encode
Data Input Queues
Local
SRAM
Decompress
Unpack/
Alignment
Result
Format/
Encode
Decompress
Unpack/
Alignment
Result
Format/
Encode
Decompress
Unpack/
Alignment
Result
Format/
Encode
Data Output QueuesAnalytics
Engine
(1 of 32 on M7)
On-Chip
Network
Data Input Queues
Data Output Queues
On-Chip
Network
On-Chip
Network
On-Chip
Network
Scan, Filter,
Join
Scan, Filter,
Join
Scan, Filter,
Join

SQL in Silicon: Accelerating Oracle Database 12c Queries
19
While Off-Loading Cores to Carry-on with other Work
One
step 10Xfaster
Decompress
More than Doubles data size
Read
Software
scan
Read
Write
Write
Read
DAX
Write
Multiple
steps
SQL:
SELECT count(*)
…WHERE lo_orderdate = d_datekey
…AND lo_partkey = 1059538
AND d_year_monthnum BETWEEN 201311 AND 201312;
t

Image
Apache Spark with Software-in-Silicon
One billion rows filtered and folded into cube
Faster
6X
With In-memory
Analytics Accelerator
Without In-memory
Analytics Accelerator
38
seconds
6 seconds
20

* Best 2-chip result
SPECjEnterprise * PeopleSoft HR+Payroll
SPECcpu rates 1-chip (four benchmarks) PeopleSoft HR-only
SPEComp PeopleSoft Batch-only
SPECvirt2013 * PeopleSoft FMS Batch
Java Workload FLEXCUBE UBS
SAP-SD Oracle Communications ASAP Telco
E-Business Online
E-Business Order-to-Cash Batch In Memory Aggregation
E-Business Payroll Batch NoSQL YCSB
Oracle Internet Directory Graph
Oracle Stream Explorer R statistics
SPARC M7: Setting 20 World Records in Performance
21

Real Time Analytics:
83X performance
improvement using T7
Database Acceleration
“Using Oracle SPARC T7 servers,
we can reach new levels of
insight with real-time queries on
up-to-date transactional data.”
Shlome Seidenfeld CIO, VP E-
Commerce, B&H Photo Video

North American Music
Company: Improve Security
SPARC Solution Benefits:
• Better Security & compliance
• 9 to 1 consolidation
• 80% lower power/cooling
• 66% increase in throughput
Solution Details
• SPARC T7 and T5 servers
• Oracle Database EE

Large Financial Company:
Scalability and Capacity
SPARC Solutions Benefits:
• Capacity beyond 5 years
• 44% IT budget reduction
• Meets SLAs
Solution Details
• Four SuperCluster M7 systems
• Oracle Database, EBS,
Hyperion, and Data Guard

SPARC Systems for Modern Enterprise Workloads
25
SPARC S7
Servers
SPARC T7
Servers
• Small-Med DB, Bus Apps
• Web tier
• Cloud Infrastructure
• *Telco/NEBs compliance
• Large-Very Large DB
• Enterprise Bus. Apps
• Consolidation
SPARC
SuperCluster M7
• Single Instance,
Clustered DB
• Simple to deploy
• Integrated
Security
• Extreme DB
Performance,
• Enterprise Bus.
Apps
• Consolidation
SPARC
MiniCluster S7SPARC M7 Servers
SPARC T7
Servers
SPARC S7
Netra S7*
Servers
Scale-out Scale-up Appliance Engineered System

Oracle Cloud Compute Services
2610/26/2016
Bare Metal Elastic
Compute
Engineered
Systems IaaS
Docker
Service
Compute
Multiple OS
Multiple
Hypervisors
Docker
Containers
Bare Metal
Elastic Compute
Dedicated
Compute

Bare Metal Elastic
Compute
Engineered
Systems IaaS
Docker
Service
Compute
Multiple OS
Multiple
Hypervisors
Docker
Containers
Bare Metal
Elastic Compute
Dedicated
Compute
27
Oracle Cloud Compute Services
10/26/2016
Dedicated
Compute
Dedicated Compute Capacity - SPARC Model 300
Dedicated
Compute
Only Oracle
offers
SPARC M7 and x86
Same Price
for
SPARC M7 and x86

SPARC Model 300 Hardware Layout
• SPARC Compute
– 300 OCPU (300 SPARC M7 Cores)
– Each compute node has 30 cores (240 Threads/vCPUs) and
480GBs of memory in a Guest Domain (Dom U) visible to the
customer
– Runs Solaris 11 (including Solaris 10 Branded Zones)
– Customer must bring their own licenses for Oracle software
• Storage
– Local (On-board) Storage in SPARC servers for OS boot images
– Integrated ZFS Storage Appliance with 32TB of usable capacity
for Virtual Machines (Solaris Zones) and application storage
28RESTRICTED CONFIDENTIAL

Cloud Bursting
• Capacity on demand
• Maximize seasonal impact
Production
• Always on, low overhead
• Low overhead Encryption
& Silicon Secured Memory
for data protection
Comprehensive
Test/Dev
• Fast, self-service VMs
• Optimize applications
using Oracle Solaris Studio
Disaster Recovery
• Fast access to secondary
sites
Dedicated SPARC Cloud Service: Use Cases
RESTRICTED CONFIDENTIAL

Customer
Data Center
Integrated Private Cloud and Public Cloud Offerings
Archive Cloud
SPARC Dedicated Compute
Oracle Cloud
Storage Backup
Oracle Confidential – Internal and Authorized OPN Partners Only
SPARC Private Cloud

Path to a Modern Infrastructure and the Cloud
31
Standardize Your
Unix systems
Deploy cloud
services
Modernize your
SPARC Systems
SPARC S7/M7

Rules • Margin Goals • Behavior Goals • Specials •
Partners • Sales Role • Order Receipt • Order Acceptance •
Assign Credit • Assess Disputes • Make Payments • Mid
Year Adjustments • Compliance Reporting • Management
Insight Reporting • Pick, Pack, Ship, Bill • Spend
Classification and Analysis • Procure to Pay • Catalog
Production • Contract Management • Job Status Inquiry •
Financial Close • Supply Chain Management • Rate Change
Impact Analysis • Instant Part Reuse • Analysis • Order
Importation • Work Order-Less Completion • Root Cause
Analysis •Business Strategy and Goals • Products • Quotas •
Payment Rate • Rules • Margin Goals • Behavior Goals •
Specials • Partners • Sales Role • Order Receipt • Order
Acceptance • Assign Credit • Assess Disputes • Make
• Do you have a plan or mandate to
move your legacy Unix applications to
a public cloud?
• Does your current infrastructure
investment have cloud insurance?
• Why SPARC? Why Not SPARC?
32Copyright © 2016, Oracle and/or its affiliates. All rights reserved. |

Innovation to grow.
Insight to control.

C5 journey to_the_cloud_with_oracle_sparc

More Related Content

What's hot (20)

Viewers also liked (20)

Similar to C5 journey to_the_cloud_with_oracle_sparc (20)

More from Dr. Wilfred Lin (Ph.D.) (19)

Recently uploaded (20)

C5 journey to_the_cloud_with_oracle_sparc