Řešení postavená na IBM System x serverech - HPC, Cloud, Business Analytics, SAP HANA, Big Data, Martina Kocourková & Milan Král, IBM STG System x Univerzita, Olomouc, 12.6.2013 1
Agenda • Úvod • IBM Technical Computing – HPC, Deep Computing • Co to je? • iDataPlex a Flex Cluster řešení • Podpůrné SW (Platform Computing, GPFS...) • Cloud/ Virtualizace/ Automatizace • Standartizace – Gartner • SmartCloud Entry • SCEntry Demo • BAO a Big Data • Klasická architektura BA • In Memory koncept • SAP HANA • Cognos/ DB2 BLU • Big Data • Závě ěr 2
IBM Technical Computing – HPC, Deep Computing
3
Hlavní atributy IBM Technical Computing řešení • • •
•
•
Technical Computing již dávno není • pouze o tzv. high-end superpočítačích Překonání tradičních bariér HPC díky řešení od IBM Podniky a organizace napříč odvětvím, geografickým umístěním a velikostí mohou díky IBM využít sílu Technical Computing řešení jako svou konkurenční výhodu Nižší náklady a vyšší produktivita díky maximálnímu využití výpočetní kapacity a výkonu Snížení rizik díky optimalizovaným „end-to-end“ řešením šitým na míru potřebám zákazníků
IBM Technical Computing řešení zahruje infrastruktutní systémy, software a služby: – IBM Platform Computing™ – IBM GPFS™ – IBM System Storage® DCS3700 – IBM Intelligent Cluster™ – IBM® System x® servers
4
IBM System x portfolio Enterprise x3950 X5, x3690 X5
BladeCenter iDataPlex Server consolidation, large virtualization and enterprise workloads
dx360 M4
Massive scalescale-out HPC, cloud, grid, energy efficiency
Scale Up
BC H, HS23, HX5 Infrastructure integration and simplification, energy efficiency
PureSystems
Scale Out
Single, infrastructure applications
x3550 M4, x3650, x3750 M4
System x Rack and Tower 5
5
iDataPlex – Navržen pro větší flexibilitu datových center Široké portfolio komponent, které se přizpůsobí výpočetním potřebám Vašich zákazníků Rack Management Appliance
2U Chassis
Server Tray Dual GPU IOTray
iDataPlex Rear Door Heat Exchanger
Storage Tray Storage Drives & Network Options
Switches (front)
PDUs (rear) 3U Chassis
6
6
IBM Flex System Enterprise Chassis Enterprise Chassis Design
Chassis
Fans
CMM
Standard Node bays 14 Node Bays (7 Full Wide)
Infrastruktura
Nejširší konektivita na trhu včetně podpory technologií Ethernet, FC, FCoE, ISCSi a IB
☺ Flexibilní architektura (flexibilní rozšiřování výpočetního výkonu i úložných kapacit)
☺ Jednoduchá integrovaná správa a nižší provozní N ☺ Nyní podporuje 40 Gb Ethernet, připraveno pro 100 Gb
Power Supplies (6X)
10 U REAR Scalable Switch Bays
• • • • • •
Integrated Integrated Storwize V7000 Storage*
FRONT
4 pozice pro switche 10U Chassis, 14 obsaditelných pozic 6 pozic pro napájecí zdroje Až 8 ventilátorů Light Path diagnostika Integrovaný chassis management modul 7
Integrace bez kompromisů, navrženo pro další dekádu Flex System x240 Compute Node
Výpoč četní výkon
Infrastruktura
Flex System x220 Compute Node
Systemové Portfolio připraveno na velké pracovní zátěže ☺ Snížení nákladů na pořízení nových serverů prostřednictvím virtualizace/ konsolidace ☺ Flexibilita nasazení a možnost integrace se stávajícím HW
Flex System x222 Compute Node* Flex System x440 Compute Node Flex System p260 Compute Node IBM 2S PowerLinux Compute Node Flex System p460 Compute Node IBM PCIe Expansion Node 8
* 2Q13 target
IBM Confidential
Platform Computing
IBM Platform Product Family Platform HPC for System x: • easy-to-use cluster management, User-friendly • GPU scheduling and reporting • Robust commercial MPI library, including IBM Platform MPI
Platform LSF: • • • •
Workload management for HPC Optimal utilization: reduced infrastructure cost High throughput: faster time to results Policy and resource-aware scheduling
Platform Symphony: Low-latency grid management platform for: • distributed computing • big data analytics with sophisticated resource sharing. Deploy standalone or with IBM’s InfoSphere BigInsights.
Platform Cluster Manager: • Quickly provisions, runs, manages and monitors HPC clusters with unprecedented ease • Single & MultiTenant
This presentation is intended for the education of IBM and Business Partner sales personnel. It should not be distributed to customers. 9 © 2013 IBM Corporation 9
Platform Computing
IBM GPFS GPFS
Overview IBM GPFS is a high-performance, distributed, parallel file system that is complementary to both IBM Platform Symphony and IBM Platform LSF
GPFS
GPFS
Key Capabilities •High-performance parallel file system •Supports multiple operating environments and interconnects •A single global namespace •Automated data lifecycle management & archival •Highly reliable storage •File-placement option optimized for Hadoop MapReduce
Maintain a single view of data enterprise-wide to promote collaboration
Local Area Network (LAN)
Client Benefits •Improve application performance and resource utilization by removing data-related bottlenecks •Supports improved collaboration across geographies by providing a global view of data with efficient remote access •Improves cost-efficiency by automatically moving infrequently accessed data to more cost-efficient storage (ie. tape) •Simplification - use a single filesystem for traditional applications as well as Hadoop MapReduce applications
GPFS
Dramatically improve file system performance to remove data bottlenecks
This presentation is intended for the education of IBM and Business Partner sales personnel. It should not be distributed to customers. 10 © 2013 IBM Corporation 10
Platform Computing
Traditional Computing - Capacity Approach Utilization
Invisible cost of delay to productive use &utilization Extended learning period for users
Extended time to fully operational system
It takes a long time to “Sweat the Assets”
Open source management software
Time in months
This presentation is intended for the education of IBM and Business Partner sales personnel. It should not be distributed to customers. © 2013 IBM Corporation 11
Platform Computing
IBM Platform Computing - Productivity Approach
Utilization Faster time to full system readiness
Better throughput & utilization
Faster time to full user productivity
Strong policies drive allocation of resources
“Sweat the Assets” much earlier
Open source management software Platform Computing
Time in months
This presentation is intended for the education of IBM and Business Partner sales personnel. It should not be distributed to customers. © 2013 IBM Corporation 12
Platform Computing
The Result: Lower Cost of Ownership 3-year management software TCO for a 32-node cluster Open source cost less, but…
Lost productivity due to integration, troubleshooting & user learning curve
Costly to integrate software components
Source: “The Hidden Cost of Open Source” white paper, Platform Computing, 2012 This presentation is intended for the education of IBM and Business Partner sales personnel. It should not be distributed to customers. © 2013 IBM Corporation 13
Platform Computing
Pro více informací... IBM Technical computing web Clusters for Dummies (brožura) Hadoop for Dummies (brožura)
This presentation is intended for the education of IBM and Business Partner sales personnel. It should not be distributed to customers. © 2013 IBM Corporation 14
Konec HPC části…
15
© 2013 IBM Corporation
Cloud & IBM Technologie
Cloud becomes top CIO technology priority in 2011 Gartner 2011 CIO priorities
Source: Gartner – Reimagining IT: The 2011 CIO Agenda 17
© 2013 IBM Corporation
TCO kalkulace pro konsolidaci a privátní cloud Známé (příp. odhadnuté) náklady: • hw, sw • hw a sw údržbu • facility, energie, prostory • lidské zdroje, ostatní OPEX
Obtížně vyčíslitelné benefity: • ‚hodnota‘ rychlosti poskytnutí služby/provisioning? • ‚hodnota‘ interního billingu, reportingu?
Business Case Summary Total CPU Cores Used CPU Cores Total CPU Sockets #Logical Servers #Physical Servers Ave.Log.Srv RIP Total RIP Capacity Total RIP Workload Ave %Utilization
Current 8,530.00 8,530.00 4,394.0 4441.00 1943.00 1,267.9 5,630,837.1 1,185,482.0 22%
Annual Operating Costs (AOC) Staff Cost Code System Software M&S Hardware Maint Space Electric Staff Depreciation Total AOC est.potential saving /yr
Today 5,623,864 0 210,832 995,135 1,617,461 0 8,447,292
One Time Costs (OTC)
BAU + Cloud 3024 3000 756 4346 378.00 974.2 4,233,750.0 1,185,482.0 28%
Blades + Cloud 2224 1947 278 4346 139.00 547.4 2,378,990.6 1,185,482.0 50%
R O Z V
Cloud 7,727,281 0 36,275 392,319 1,455,715 0 9,611,590 -1,164,298
Fewer servers 7,405,822 0 4,288 260,351 1,310,143 0 8,980,604 -533,312
CloudBurst 1638 1561 273 4346 12.07 547.8 2,380,525.0 1,185,482.0 50%
Change -81% -82% -94% -2% -99% -57% -58% 0% 124%
3 Year Projection Millions
Sizing
45
Transition
40
Hardware Maint
35 Space 30 Electric 25
CloudBurst 1,214,658 0 11,696 263,086 1,179,129 0 2,668,569 5,778,723
Software Purchase Hardware Purchase Transition Total OTC Write-Off Net Cash Investment
0 13,529,500 0 13,529,500 0
0 2,835,000 0 2,835,000 0 -10,694,500
0 3,118,635 0 3,118,635 0 -10,410,865
0 6,818,058 0 6,818,058 0 -6,711,442
OTC + 3x AOC 3 yr saving Payback Period
38,871,376
31,669,770 7,201,606 9yr2m
30,060,448 8,810,928 19yr6m
14,823,765 24,047,611 -2yr10m
-78% 0% -94% -74% -27% 0% -68%
Depreciation
20 15
Staff
10
Software Purchase
5 Hardware Purchase 0 C
3
2
1
System Software M&S
3 Year Projection
18
Project Time 0yr0m
© 2013 IBM Corporation
Základní „IBM produkty“ v oblasti budování Cloud Computingu
Foundation
Services
Solutions
IBM Smart Cloud Entry Technologie pro vybudování 19 Cloudu
Platforma IBM Datových Center nabízející služby Veřřejného (Public) Cloudu
Software jako služžba © 2013 IBM Corporation
Smart Cloud Entry: ‚Entry‘ řešení cloudu
Verze 3.0 červen 2013 Samoobslužný portál Katalog služeb Workfow Provisioning a expirace Měření a reportování Intuitivní rozhraní Snadno nasaditelné Privátní cloud VMware KVM a PowerVM + nově HyperV 20
Demo v IBM Inovation Centru, Praha Chodov
© 2013 IBM Corporation
Závěr Cloud části…
21
© 2013 IBM Corporation
BAO: In-Memory, HANA, Big Data
Klasická architektura In Memory koncept In Memory řešení SAP HANA BigData MapReduce/Hadoop/BigInsight InfoSphere Streams
BigData Demo v IBM IIC nebo TeC v přípravě
CIO #1 Concern Business Analytics
83% 76%
Virtualization 71%
Risk Management & Compliance Mobility Solutions
68%
Customer & Partner Collaboration
68%
Self-service Portals Application Harmonization
64%
Business Process Management
64%
SOA / Web Services Unified Communications
23 23
66%
Source: IBM Global CIO Study, n = 2345
61% 60%
Komponenty
Sizing!!! Architektura Data Warehouse
Cubing Services Agregace Indexes
ETL
Operational Source Systems Structured/ Unstructured Data
24
Implementation Services
DB2 Utilities Suite
© 2013 IBM Corporation
In-Memory koncept: Rychlost RAM, SSD a pevných disků Non-Persistent, Volatile Processors
Memory
Very, very, very, very, very fast
Very, very, very fast
< 10’s ns
~100 ns
Persistent, Non-Volatile Disk
SSD
Fast
~200,000 ns
Very, very slow comparatively
1,000,000 8,000,000 ns
In-Memory DBPřístupová rychlost Klasická DB
~1 second
~33 minutes
~ 12.5 hours
Srozumitelnější pro člověka…
© 2012 IBM Corporation
25
IBM In Memory řešení Cognos TM1 IBM SAP HANA Appliance IBM DB2 BLU IBM Informix In-Memory … a další real-time calculations, what-if and ad-hoc analysis, data refresh and write-back In-memory: the whole multidimensional database runs in the physical memory so everything is incredibly fast Real-time both data refresh and calculations run real-time 26
© 2013 IBM Corporation
IBM Systémy pro koncept SAP HANA velikosti/stavební bloky/možžnosti rozšířření Scale-up… ažž na XXL (další slide) Size
XS
S
S+
M
L
Building Block
x3690 X5
x3690 X5
x3950 X5
x3950 X5
x3950 X5
Part Number (Sys x - config)
7147-HAx
7147-HBx
7143 - HAx
7143 - HBx
7143 – HBx + 7143 – HCx
Intel CPU
2 x E7 2870
2 x E7 2870
2 x E7-8870
4 x E7-8870
8 x E7-8870
RAM
128 GB DDR3 (8 x 16GB)
256 GB DDR3 (16 x 16GB)
256 GB DDR3 (16 x 16 GB)
512 GB DDR3 (32 x 16 GB)
1 TB DDR3 (64 x 16 GB)
Log Storage
10 x 200 GB 1.8‘‘ MLC SSD
10 x 200 GB 1.8‘‘ MLC SSD
1.2 TB ioDrive2 FusionIO
1.2 TB ioDrive2 FusionIO
2 x 1.2 TB ioDrive2 FusionIO
Data Storage
-
-
8 x 900 GB 10k SAS HDD
8 x 900 GB 10k SAS HDD
16 x 900 GB 10k SAS HDD
Storage Ctrl
2 x M5015
2 x M5015
1 x M5015
1 x M5015
2 x M5015
Data / Log Storage Summary
1.6 TB RAID 5 data and log storage
1.6 TB RAID 5 data and log storage
5.4 TB RAID5 data storage 1 TB log storage
5.4 TB RAID5 data storage 1 TB log storage
10.8 TB RAID5 data storage 2 TB log storage
Ethernet
4 x 10 GbE, 6 x 1 GbE
4 x 10 GbE, 6 x 1 GbE
4 x 10 GbE, 6 x 1 GbE
4 x 10 GbE, 6 x 1 GbE
8x 10 GbE, 12 x 1 GbE
Upgrade Option Scale Up or Scale out
XS - > S
Up to 16 node scale out with HA, SAP certified
S+
M L (with L-option) Up to 56 node scale out with HA, SAP certified
L XL or XXL Up to 56 node scale out with HA, SAP certified
Software
Preload: SLES4SAP GPFS SAP HANA*
Preload: SLES4SAP GPFS SAP HANA
Preload: SLES4SAP GPFS SAP HANA
Preload: SLES4SAP GPFS SAP HANA
Preload: Addtl. GPFS lic.
27
M
Scale-out… ažž 16x node S nebo 56x M/L
© 2013 IBM Corporation
IBM Systems solution for SAP HANA Scale-up pokrač čování
28
Size
XL
XXL
Building Block
x3950 X5
x3950 X5
Part Number (Sys x - config)
tbd
tbd
Intel CPU
8 x E7-8870
8 x E7-8870
RAM
2 TB DDR3 (128 x 16 GB or 64 x 32 GB)
4 TB DDR3 (128 x 32 GB)
Log Storage
2 x 1.2 TB ioDrive2 FusionIO
4 x 1.2 TB ioDrive2 FusionIO
Data Storage
16 x 900 GB 10k SAS HDD
24 x 900 GB 10k SAS HDD
Storage Ctrl
2 x M5015
2 x M5015, 1 x M5025
Data / Log Storage Summary
10.8 TB RAID5 data storage 2 TB log storage
16.2 TB RAID5 data storage 4 TB log storage
Ethernet
8 x 10 GbE, 12 x 1 GbE
8x 10 GbE, 12 x 1 GbE
Upgrade Option Scale Up or Scale out
XL XXL (with additional memory, ioDrive2, M5025, EXP 2524)
Software
Preload: SLES4SAP GPFS SAP HANA
Preload: SLES4SAP GPFS SAP HANA
© 2013 IBM Corporation
InfoSphere Streams: In-Motion Vs Traditional Analytics Stream Analytics RTAP Non-Traditional / Non- Relational Data Sources
Ultra Low Latency Results
In-Motion Analytics
Audio, Video, emails…
OLAP / OLTP Traditional / Relational Data Sources
Traditional Analytics
(Alpha)Numeric, text…
Warehouse
Streams DOES NOT store data for analysis Traditional data is finite, saved and known. Streaming data is NONE of these
Results
Streams: Industry Focus
Government
Telco
•Security, foreign intelligence, foreign surveillance and cyber security •Complex analysis of data in motion such as audio, video, email etc. •Structured and unstructured data that require highly time sensitive •Security is the #1 spending initiative for the US Gov’t.
•Real time insights into network and subscriber behavior •Improve network performance, reduce customer churn, •Increase loyalty & appeal, prevent spam, •Focus on countries w/ emerging growth leapfrogging hw phone lines •Subscriber base doubling every 2 years.
FSS Algorithmic & high frequency trading, fraud detection in real time .Trading companies reducing costs as well as to add more value to their offerings. Brokerage firms looking to find ways to add significant value to their customers
Energy / Utilities Healthcare •Smart meters, sensors, smart grid •Analysis in real time for energy savings and prevention of outages. •Analytics and complex logic & the need for extreme speed to take action. •Significant ROI is emerging.
•Real time predictive modeling & fast analysis on “Big Data •Analytics on large volumes of physiological data from increasingly sophisticated sensors and medical devices. ie:Patient Monitoring in ICU
BIG DATA: Hadoop Map-Reduce BigInsight • ‚Stream‘ dat • Nestrukturovaná data • Strukturovaná data
BIG DATA: Příklad infrastrukturní architektury
• Paralelní úlohy • HDFS • GPFS • inter-node traffic
32
© 2013 IBM Corporation
Konec BAO části
Otázky?
Děkujeme za pozornost
[email protected] [email protected] 34
© 2013 IBM Corporation