#HPEDiscover
Please give me your feedback
Session ID: T4406
Speakers: Kannan Mani and Amel Blouza
–Use the mobile app to complete a session survey
1.
Access “My schedule”
2.
Click on the session detail page
3.
Scroll down to “Feedback”
– If the session is not on your schedule, just find it via the app’s “Session Schedule” menu, click on this session and scroll down to “Feedback”
– If you don’t have it, download the event app today. Go to your phone’s app store and search for “HPE Discover 2017”
–To access the session survey online, go to the Agenda Builder in the event content catalog and click
on your session
Thank you for providing your feedback, which helps us enhance content for future events.
2
Accelerating data and analytics
with Oracle and SQL Server on
HPE infrastructure
Kannan Mani and Amel Blouza
#HPEDiscover
Presenters
Kannan Mani
WW Sr. Principal Solution Manager/Architect
Data Platforms, Hybrid IT
Amel Blouza
Solutions Architect
4
The modern, data-driven enterprise organization
Data
People
Insights
Leverages all relevant data
integrated across apps
Executives to line employees informed
by all relevant data
Predictive analytics—proactive
analytics apps on every platform
Achieves superior business outcomes
5
Inhibitors to becoming a data-driven organization
Database management is the cornerstone of data-centric foundations
Legacy architectures and infrastructure can’t handle the volume, velocity or variety of today’s data
55%
of decision makers lack complete
information1
51%
of decision makers don’t receive the
required information on time1
Ideas
24%
of organizations use data for
decision making all the time1
Insights
Data
Major factors hindering the use of data in decision making
Rigid, single-purpose infrastructures exist where mixed environments and workloads are needed
1
"Delivering Data at the Speed of Business." Database Trends and Applications, April 2016.
6
Challenges in data management
Escalating complexity
High costs
Technology limitations
Database landscape
Management
Aging infrastructure
Data growth
Software licensing
Scalability
Reduced performance
Missed opportunities
Availability
7
Right-sizing data management
infrastructure
8
Empower business processing
Align infrastructure with demanding applications
Users / transactions
Criticality
Mission-critical
Headroom / growth
Database size
Up to 1-3 TB
Up to 10 TB
Up to 30+ TB
Business processing
Find the Right Mix
–
Compute-intensive OLTP workloads
–
Memory and I/O scalability for mixed transaction/analytics workloads
–
ERP, CRM, batch …
–
Better OLTP per-core performance lowers compute, licensing needs
–
Dynamic demand, growth
–
Reduced management costs with a consolidated landscape
Optimizing infrastructure
DL380 + Persistent Memory
Virtualization
DL580 / SY680
Virtualization, consolidation
Superdome X / MC990 X
Virtualization, consolidate all tiers and mixed landscapes with flexible hard partitions
3PAR StoreServ
From midrange to high-end with multi-tenancy and built-in protection to consolidate mixed workloads
9
Quantify Oracle bottlenecks with Database Performance Profiler
Example
Objective assessment tool
By eliminating cores, this customer could save cost on Oracle licensing fees1
– DPP report provides insights and trends about your workload
– Chart below generated on a 64-core server
– CPU utilization never goes over 12%
Where’s the slowdown?
Storage, servers or both, DPP
will show you!
– Workload could be converted to a new, smaller server
– A new efficient ProLiant server or a blade in a Superdome X is effectively free
1,
Based on an example in Technical white paper, “HPE Reference Architecture for Oracle 12c license savings with HPE 3PAR StoreServ All Flash and ProLiant DL380 Gen9,” Feb 2017
10
Customer lessons for data management
EMIS Healthcare
11
EMIS Healthcare
Customer lessons for data management
Leading health sector with innovative technologies,
from market-leading clinical management systems to
health analytics tools James Atkinson
Solutions Design Manager,
Hosted Services
12
Overcoming data management challenges
Costs, complexity, performance
13
Meeting performance needs and cutting license costs
– Server optimization / consolidation
– Scale-up efficiency / Reduce cluster latency
– Better per-core performance
– Avoid Oracle RAC
Database migration?
Oracle costs higher
comparing cost of SQL Server
In-DB advanced analytics
– HPE Persistent Memory
– Accelerate database performance with All-Flash storage
– Reduce latency and improve back-end performance
– Scale workload to hundreds of thousands of IOPs
Mobile BI & self-service BI
Everything
built-in
#1 TPC-H—DW
Built-in ETL
Industry leader—OLTP
– In-memory database technology
Multiple features adds
up to higher license cost!
For OLTP & DW scenario, price comparison is based on a server with 2 proc, 8 cores each
14
Server optimization to control costs and complexity
Consolidating servers and database instances
1996-2018 Worldwide Spending:
New Server, Power and Cooling, and Management and Administration
– Simplify database landscape and management
– Consolidate licensing
– Potentially Increase utilization of Database/App licenses
(Depends on degree of consolidation)
– Reduce hardware costs > 50%1
– Consolidate servers based on workload, avoid “Server Sprawl”
(Depends on the environment)
– Lower hardware/software support costs
– IDC Over 35% in savings from consolidating to scale-up servers
– Increase IT staff productivity
– Better resource utilization (CPU, memory, network)
“The shift to low-capex scale-out x86 systems also had undesirable
effects. The increased opex of managing the expanding x86-installed
base has been taxing IT budgets, making it difficult to fund new
initiatives that might have delivered revenue-generating innovations.”
1,
Based on example in Technical white paper, “HPE Reference Architecture for
– Lower software licensing costs
– Reduce energy consumption
– Reduced IT infrastructure costs
– Reduced end-user productivity losses
Microsoft SQL Server 2016 on HPE Integrity Superdome X and HPE 3PAR StoreServ 8450 ,” Aug
2016
15
Server optimization—from scale-out to scale-up
Drive up performance while you drive down license fees1
For typical OLTP and mixed transaction/decision support database workloads, a scale-out
cluster requires more cores than a scale-up server, to provide equivalent performance.
– Scale-up benefits
Eliminate cluster latency,
Oracle RAC fees
3-year TCO model,
figures in $K USD
$18,000
– Support very large contiguous memory to accelerate in-memory databases
– Offer good performance without significant tuning
– Eliminates need for partitioning for scale
38%
$12,000
lower TCO
– Easier to deploy and manage than scale-out
– Higher hardware costs per node, but fewer nodes
– Deliver redundancy/failover for highest levels of availability with just two nodes
– Can be clustered for very large workloads
$6,000
Other
Software
Support
Hardware
Support
Software
Hardware
– Consider scale-out compute for
– Better performance for select analytics workloads
– Lower cost of initial and subsequent server hardware acquisition
$0
Exadata
Superdome X
– But offset by higher database licensing costs for medium to larger implementations and lower per node utilization
– Addition/removal of servers by partitioning compute at node level
– Implementing rolling upgrades
1,
Based on HPE Internal study, “HPE Reference Architecture for Oracle 12c high availability with HPE Integrity Superdome X, 3PAR StoreServ 8450 All Flash and Serviceguard for Linux ,” Mar 2017
16
Boosting performance with HPE Persistent Memory
Highest performing persistent memory in the market running at DRAM speeds1
New HPE Scalable
Persistent Memory
HPE NVDIMMs
Oracle + Persistent Memory
Up to
50%
Reduction in database licensing costs using the fastest tier of storage
on HPE servers2
Up to
4X
Increased Oracle database throughput with HPE Persistent Memory
as compared to 15K RPM SAS drives with only a 40% price increase3
Up to
3X
More cost effective than SSDs3
In-memory compute
HTAP – Real-Time Analytics
Checkpoints and restores
Up to 20x reduction in database
recovery time preserving maximum
uptime.
Up to 27x faster checkpoint
operations enabling significantly
faster business operations
TB-scale databases
SQL Server + Persistent Memory
– Double the performance and half the latency using byte-addressable storage
features in SQL Server 2016
OLTP
+
OLAP
=
HTAP
And much ... much … more
– A single HPE NVDIMM doubles SQL Server OLTP performance
– Gains ranged from 2-4 X depending on the baseline tested against
– Manage SQL Server 2016 performance outside of Persistent Memory with 3PAR
AFA continuous data management
Big Data
Service Providers Performance Tier Virtualization
1,2,
3,
Based on internal testing performed by HPE
Technical white paper, “Improving Oracle Database performance with HPE Persistent Memory on HPE ProLiant DL380 Gen9,” August 2016
17
In-memory database for real-time business insights
When the database is loaded in system memory, it’s lightning fast
HTAP
OLTP Transaction
OLAP Analytics
SQL Server in-memory
Accelerate insights and transactions
Faster Transactions
In-memory OLTP
Up to 30x faster transaction processing
with In-Memory OLTP
Row-based
in-memory database
Column-based
in-memory database
Faster Queries
In-memory data warehouse
Over 100x query speed and significant
data compression with
In-Memory ColumnStore
18
One stop solution for all of your enterprise needs
Server on demand
–
HC380/SimpliVity and
Synergy Image streamer
through HPE OV
–
Provision databases on
demand
–
–
Minutes to provision in
production and in the
dev/test environment
Automated Migration from
Physical to Virtual
QoS (HA and DR)
–
All of our platforms
(DLs, BLs, Hyper
Converged, MCS
(Superdome/MC990x),
Synergy and
3PAR/Nimble)
–
Increase application/DB
Quality of Service
–
Scale dynamically
–
Simplified high availability
and simple disaster
recovery
Security
–
Next generation of servers
are more secured with
HPE iLO
–
Complete isolation
between Production and
Test/Dev systems
–
Support
–
HPE Data care Multivendor support include
Oracle and SQL Server
support
–
One vendor to deal for
support
Protects databases and
applications against
network-based threats
19
Industry’s broadest portfolio for enterprise database apps
Selected compute – 100% cloud ready
ProLiant DL380 with
Persistent Memory
Hyper Converged 380
SimpliVity
ProLiant DL580 and
Synergy SY680
Cost-effective 2-socket server;
boost OLTP with Persistent
Memory
“VM vending machine,”
consolidate multiple virtualized
databases
Scale-up 4-socket solutions for
business processing, analytics and
consolidation
4X faster OLTP on
SQL Server 20161
Simple to deploy,
ready for hybrid cloud
High availability, 2-4 sockets,
6 TB memory
HPE Superdome Flex
Integrity Superdome X
Mission-critical x86 solutions for large enterprise
applications and data center optimization; up to 48 TB memory for inmemory database and analytics; 99.999% single system availability
Flexible modular scalability,
from 4 to 32 sockets
Up to a 4x increase in Microsoft SQL Server 2016 transaction performance with HPE ProLiant DL380 Gen9 and HPE Persistent Memory, as compared to SAS SSDs.
Based on Internal HPE lab testing on HPE ProLiant DL380 Gen9 E5 2600 v4 with HPE 8GB NVDIMM-N as compared to SAS SSDs, June 2016.
1-8 server blades,
2-16 sockets,
1
20
Industry’s broadest portfolio for enterprise database apps
Selected storage, networking
3PAR StoreServ 9450
All Flash Array
3PAR StoreServ 20000
All Flash Array
Lightning fast IO and built-in resiliency. One architecture from midsize to
high-end, eliminating tier 1 distinctions, with “six-nines” availability, 75%
data reduction for affordability.
2 million IOPS,
6PB RAW
3.7 million IOPS,
228 SSDs per rack
StoreOnce Cloudbank
Rapid app-integrated data
recovery, transparent backup with
Recovery Manager Central
Use any application,
leverage any cloud
Arista 7300 & 7150
Arista 7500R, 7280R
and 7020R
Ultra-low latency networking for enterprise-class database deployments.
Leading solutions for high performance big data clusters, 100GbE cloud
data centers, multi-rate uplinks.
SSD options,
large forwarding tables
FlexRoute routing solutions;
manage by CloudVision
21
HPE Data Management Reference Architectures
– HPE Reference Configuration for Microsoft SQL Server 2016 on HPE Synergy composabl
e infrastructure
– HPE Reference Architecture for Oracle 12c license savings with HPE 3PAR StoreServ A
ll Flash and ProLiant DL380 Gen9
– HPE Reference Architecture for deploying Oracle 12c with HPE Synergy Image Streamer
– HPE Reference Architecture for HPE ProLiant DL580 Gen9 and HPE 3PAR StoreServ 8450:
Consolidating multiple Oracle databases into a single instance Oracle database
– Improving Oracle Database performance with HPE Persistent Memory on HPE ProLian
t DL380 Gen9
– HPE Reference Architecture for Microsoft SQL Server 2016 Operational Analytics on
the HPE ProLiant DL580 Gen9
– HPE Reference Architecture for Microsoft SQL Server 2014 on HPE 3PAR Storage with D
Check out all Data Management Reference
Architectures www.hpe.com/info/dm-RA
isaster Recovery
22
Get more information
Attend these sessions
Visit these demos
Wednesday November 29
–
12:30 – 13:30
Business acceleration: Compute innovations driving insights and intelligence
to outpace your competition (SL4410)
DEMO 1101 HPE’s new mission-critical x86 platform delivers a
breakthrough design and unparalleled compute power for real-time
analytics
–
14:00 – 15:00
Announcing HPE’s next-generation Superdome platform – Unparalleled
flexibility and scale for real-time analytics (B4381)
DEMO1102 Empower a data-driven enterprise with SAP HANA:
Accelerate your real-time business transformation with HPE HANA
portfolio
–
DEMO1103 Stephen Hawking expands understanding of our universe
with next generation in-memory compute
–
DEMO1104 German research institute leverages Memory-Driven
Computing to fight global time bomb of neurodegenerative diseases
17:30 – 18:00
Accelerate your real-time business—as In-memory databases race into the
mainstream, are you ready? (DF4386)
Thursday November 30
9:30 – 10:30
Now is the time to act; accelerate your transformation to SAP HANA with the
right mix of HPE infrastructure and services (B4380)
If you have a CDA, visit the Technology Showcase to discuss future technology
12:30 – 13:30
The
of HPE mission-criticalforcompute:
Cloud-ready solutions that excel
Visitfuture
www.hpe.com/superdome
more information
in a Hybrid IT environment (B4382)
23
Please give me your feedback
Session ID: T4406
Speakers: Kannan Mani and Amel Blouza
–Use the mobile app to complete a session survey
1.
Access “My schedule”
2.
Click on the session detail page
3.
Scroll down to “Feedback”
– If the session is not on your schedule, just find it via the app’s “Session Schedule” menu, click on this session and scroll down to “Feedback”
– If you don’t have it, download the event app today. Go to your phone’s app store and search for “HPE Discover 2017”
–To access the session survey online, go to the Agenda Builder in the event content catalog and click
on your session
Thank you for providing your feedback, which helps us enhance content for future events.
24
Thank you
Kannan.mani@hpe.com
Amel.blouza@hpe.com
25
Download PDF
Similar pages