4. Agenda
– Test Information
– Exam Objectives
– Disk Products
– Tape Products
– Tools
– Software
– Tivoli
Partnering for value and growth
ARROW ECS Confidential
5. Test 000-105: Power Systems with POWER7 and AIX
Sales
• Test information:
• Number of questions: 40
• Time allowed in minutes: 60
• Required passing score: 67%
Partnering for value and growth
ARROW ECS Confidential
8. IBM leads the UNIX Server Market
0%
5%
10%
15%
20%
25%
30%
35%
40%
45%
50%
3Q99
1Q00
2Q00
3Q00
4Q00
1Q01
2Q01
3Q01
4Q01
1Q02
2Q02
3Q02
4Q02
1Q03
2Q03
3Q03
4Q03
1Q04
2Q04
3Q04
4Q04
1Q05
2Q05
3Q05
4Q05
1Q06
2Q06
3Q06
4Q06
1Q07
2Q07
3Q07
4Q07
1Q08
2Q08
3Q08
4Q08
1Q09
2Q09
3Q09
4Q09
1Q10
2Q10
3Q10
4Q10
1Q11
2Q11
HP Sun/Oracle IBM Other
IDC Tracker: WW UNIX Server Revenue Share – Rolling 4Q Average
A record 800+ migrations to Power in 2010 (over 90% from Solaris and HP/UX)
9. IBM POWER Systems
Complete flexibility for workload deployment
– Extensive range
– Binary compatibility
– Mainframe-inspired reliability
– Hypervisor-based virtualisation
– AIX, Linux and IBM i OS
770/780
795
Enterprise
710/730
PS701/2
PS700
720/740
Express
775
755
HPC
IBM
750
PS703/4
Power775supercomputer
Power780
10. Power795
The Benefits of Virtualisation
– Shared processors
– Increased server utilisation
– Potential reduction in s/w costs
– Shared adapters
– Reduction in server costs
– Reduction in infrastructure costs
– Reduced environmental impact
– Savings in heat, power and space
– Rapid service provisioning
– Reduction in time for procurement,
provisioning and installation
Shared Network
Adapters
Shared Disk
Adapters
Shared Process Pool
Fewer switches
and routers
11. Power vs Linux and Intel
– Individually, x86 systems are cheaper than POWER7 systems
– But … multiplication of x86 servers inflates costs for
– Software licensing
– Support
– Tools for systems management, security and clustering
– SAN switches, Network routers
– Energy and space
– VMware, in practice, has limited effectiveness
– Could you run your entire workload on one system if you had a sufficiently
powerful server?
– If not, why not?
12. Linux – a matter of perception
Medium Large V. Large
Organisation Size
End User Satisfaction*
AIX Linux
*Calculated as the inverse of user complaints relative to the results for Windows
Adapted from “Does your OS
Matter?” a survey of over 43,260
environments by Solitaire Interglobal
Ltd. October 2011
13. AIX vs Linux - TCO and Business Agility
– AIX users report
– Lower staffing costs overall (due to
tools, stability, etc.)
– Lower datacentre costs
(environmental, facility, etc.)
– More highly-leveraged platforms
– …and a definite agility advantage
due to
– Ability to easily shift resources to
accommodate new
implementations
– Robust toolset for management
0
0.5
1
1.5
2
2.5
CostperFunctionPoint
Medium Large V. Large
Organisation Size
Total Cost of Ownership
AIX Linux
0
1
2
3
4
5
6
7
8
9
10
Wallclockdays
Medium Large V. Large
Organisation Size
Time to implementation
AIX Linux
Adapted from “Does your OS Matter?” - Solitaire Inerglobal Ltd. October 2011
14. – Unmatched reliability for UNIX servers
– Dynamic de-allocation of failing components
(e.g. processors and cache)
– Chipkill memory technology
– Alternate Processor Recovery
– First Failure Data Capture
– Concurrent firmware upgrades
– Logical Partition isolation
– + 40 years of innovation in large systems
– Hot-pluggable processor nodes / books
– Bulk Power Assemblies
– Redundant service processors
– Active Memory Mirroring for Hypervisor
Power795
Reliability by Design
16. The IBM POWER7 Chip
32 MB Shared L3 (custom eDRAM)
Local Region
MemoryController
MemoryController
Buses to local chips
P7
Core
256KB L2
Buses to remote chips and I/O
P7
Core
256KB L2
P7
Core
256KB L2
P7
Core
256KB L2
Local Region Local Region Local Region
Local Region Local Region Local Region Local Region
P7
Core
256KB L2
P7
Core
256KB L2
P7
Core
256KB L2
P7
Core
256KB L2
“POWER7 is a textbook on the lessons designers will
need to study in coming years” - Ron Wilson, Executive
Editor -- EDN, 8/27/2009
– Intelligent design
– Intelligent Threads
– Dynamically selects 1/2/4 per core
– Intelligent cache
– Idle cores donate cache to others
– EnergyScaleTM technology
– Dynamic frequency scaling (DFS)
– Adjusts GHz to conserve energy or boost
performance
– TurboCore mode
– Combines DFS and Intelligent Cache
– Switches off half the cores/chip
– Boosts the clock speed by ~10%
– Leaves all 32MB cache active
17. Single-socket, entry blade
One socket (single width)
4-core @ 3.0 GHz
Up to 64 GB DDR3 memory
Two SAS/SSD disks
Two integrated 1Gb Ethernet ports
Two PCIe expansion slots
Supported in BCE-R, S, T, H and HT
chassis
PS700 Blade (8406-70Y)
18. Choice of two or four sockets
Single socket (single width)
8-core @ 3.0GHz
Up to 128 GB DDR3 memory
One SAS/SSD disk
Two integrated 1Gb Ethernet ports
Two PCIe expansion slots
Two socket (double width)
16-core @ 3.0GHz
Up to 256 GB DDR3 memory
Two SAS disks
Four integrated 1Gb Ethernet ports
Four PCIe expansion slots
Supported in S, H and HT chassis
PS701/PS702 Blades (8406-71Y)
19. Choice of two or four sockets
Single socket (single width)
16-core @ 2.4 GHz
Up to 256 GB DDR3 memory
One SAS disk or two SSDs
Two integrated 1Gb Ethernet ports
Two PCIe expansion slots
Two socket (double width)
32-core @ 2.4 GHz
Up to 512 GB DDR3 memory
Two SAS disks or four SSDs
Four integrated 1Gb Ethernet ports
Four PCIe expansion slots
Supported in S, H and HT chassis
PS703/PS704 Blades (7891-73X/74X)
20. Power 710 (8231-E1C)
– 2U rack-optimised server
– 4-core @ 3.0
– 6-core @ 3.7 GHz
– 8-core @ 3.55 GHz
– Up to 128 GB memory
– I/O options
– Either 6 SFF disks or 3 SFF disks + tape
– Integrated RAID10 support
– 2-port 10Gb Ethernet adapter
– Five PCIe Gen2 x8 slots
– Not hot swap
– No I/O drawers
– 3-year warranty
IBM
21. Power 730 (8231-E2C)
– 2U rack-optimised server
– 8-core @ 3.0 or 3.7 GHz
– 12-core @ 3.7 GHz
– 16-core @ 3.55 GHz
– Up to 256 GB memory
– I/O options
– Either 6 SFF disks or 3 SFF disks + tape
– Integrated RAID10 support
– 2-port 10Gb Ethernet adapter
– Five PCIe Gen2 x8 slots
– Not hot swap
– Up to two I/O drawers
– Uses one PCIe slo
IBM
22. Single-socket server
4U rack or desk-side packaging
4,6 or 8 core POWER7 @ 3.0 GHz
Up to 256 GB memory
Up to 8 SFF SAS disks with optional RAID
External 4-channel SAS port, split backplane
2-port 10Gb Ethernet adapter
Up to nine PCIe Gen2 slots
Not hot-swap
Four low-profile slots (opt)
Expansion
12X I/O drawers
3 year Warranty
Power 720 (8202-E4C)
23. Two socket server
4U rack packaging
Up to 16-core POWER7
4/8c @ 3.3 or 3.7 GHz (4-core SCM)
6/12c @ 3.7 GHz (6-core SCM)
16c @ 3.55 GHz (8-core SCM)
Up to 512 GB memory
Up to 8 SFF SAS disks with RAID
2-port 10Gb Ethernet adapter
Up to nine PCIe Gen2 slots
Not hot swap
Four low-profile (opt)
Expansion
12X I/O drawers
3 year Warranty
Power 740 (8205-E6C)
24. Power 750 (8233-E8B)
– Industry-leading 4-socket server
– Up to 32-core POWER7
– 4-16c @ 3.72 GHz (4-core SCM)
– 6-24c @ 3.72 GHz (6-core SCM)
– 8-32c @ 3.22, 3.61 (8-core SCM)
– Up to 512 GB memory
– Up to 8 SFF SAS disks with optional RAID
– External SAS port
– Integrated Virtual Ethernet
– Five expansion slots
– Two PCI-X
– Three PCI-e
– Expansion
– PCI-X and PCI-e I/O drawers
25. LinuxAIX
Ethernet
Driver
Ethernet
Driver
AIX
Ethernet
Driver
Logical ports
Physical ports
GX Bus
H
E
A
Timeslicing
Integrated Virtual Ethernet
– Host Ethernet Adapter
– High Performance
– Uses system bus protocols instead of PCI-
X or PCI-e
– State-of-the-art packet acceleration
– Self-virtualising
– Provides Virtual Ethernet without the need
for a VIOS
– Can be shared by up to 32 LPARs
– Integrated Layer-2 switch
– Provides low latency interconnect between
LPARs
– Maximum of two 10Gb ports or four 1Gb
ports per HEA
26. IBM Watson
– Breakthrough technology for
– Natural language processing
– Unstructured text analytics
– Hardware:
– Power 750 servers + IBM storage
– Software:
– IBM Content Analytics
– Info Sphere BigInsights
– Applications:
– Medical Diagnosis, Patent
searches, Legal precedents …
Watson beat Jeopardy! world champions Ken Jennings and Brad Rutter in
a two-game exhibition contest aired on February 14 – 16, 2011
27. Unstructured data is complex
On 27th May 1498, Vasco da Gama
landed in Kappad Beach
On 27th May 1498, Vasco da Gama
landed in Kappad Beach
celebrated
May 1898 400th anniversary
arrival in
In May 1898 Portugal celebrated
the 400th anniversary of this
explorer’s arrival in India.
Portugal
27th May 1498
Vasco da Gama
Temporal
Reasoning
Statistical
Paraphrasing
GeoSpatial
Reasoning
explorer
On 27th May 1498, Vasco da Gama landed
in Kappad BeachOn the 27th of May 1498, Vasco da
Gama landed in Kappad Beach
Kappad BeachIndia
• Search Far and Wide
• Explore many hypotheses
• Find & judge evidence
• Many inference algorithms
landed in
28. The IBM POWER 755 (8236-E8C)
– 4U rack server for clustered applications
– 32-core POWER7 @ 3.61 GHz
– 128-256 GB memory
– Up to 8 SFF SAS disks with optional RAID
– External SAS port
– Integrated Virtual Ethernet
– Quad 1Gb or Dual 10Gb
– Five expansion slots
– Two PCI-X
– Three PCI-e
– No PowerVM
– No I/O drawers
– No support for IBM i
29. The IBM Power 770 (9117-MMC)
– Modular System, up to 64 cores
– Choice of configurations
– 6-48c @ 3.72 GHz
– 8-64c @ 3.3 GHz
– Up to 4 TB DDR3 memory
– Each node:
– Up to 6 SFF SAS disks with RAID
– Dual controllers, split backplane support
– Six PCIe Gen2 slots
– Four Ethernet ports
– Two 10Gb + Two 1Gb
– Active Memory Mirroring (opt)
30. The IBM Power 780 (9179-MHC)
– Modular System, up to 96 cores
– 4-64c @ 3.92 GHz (MaxCore mode)
– 4-32c @ 4.14 GHz (TurboCore mode)
– 4-96c @ 3.44 GHz
– Up to 4 TB DDR3 memory
– Each node:
– Up to 6 SFF SAS disks with RAID
– Dual controllers, split backplane support
– Six PCIe Gen2 slots
– Four Ethernet ports
– Two 10Gb + Two 1Gb
– Active Memory Mirroring (std)
– 24x7 Warranty + PowerCare
Power780
31. SAP 2-Tier Performance
– Fewer processors to do the work
– Lower software costs
– Lower costs for power / cooling
– Lower maintenance costs
– Lower cost of acquisition
0
0.2
0.4
0.6
0.8
1
SAPs SAPs/core
64c POWER7 64c Nehalem-EX
~2x more performance per core!
12-Oct-11
32. Power780
Reliability by Design
Memory ...
Chipkill with spare
memory modules
Processors ...
Alternate Processor
Recovery
Disks...
Add, remove,
replace HOT
Nodes …
HOT add, cold
repair
Adapters...
Blind-swap
PCI-e
Power...
Fans...
Remove,
replace HOT
First Failure Data
Capture ...
Virtually eliminates
intermittent failures
33. Solid State Disks
– eMLC Technology
– Combines enterprise
performance and reliability with
the cost effectiveness of MLC
Flash technology
– PCIe SAS Adapter
– Double-wide card
– 4 SSD bays + SAS Controller
– Support for RAID 0, 10, 5
– Up to 708 GB per card
– Supported on all 19” rack
models
– Standalone modules for Power
Blades and PCIe I/O drawers
177GB
SSD
SAS
Cntrl
177GB
SSD
177GB
SSD
177GB
SSD
34. 19” PCI-e/DASD I/O Drawer (#5802/#5877)
4U rack-mount expansion unit
Ten blind-swap PCI-e slots
Up to 18 hot-swap SFF disks
Selectable as 1, 2 or 4 groups
Requires use of PCI-e slots (no internal adapters)
Redundant power supplies
Diskless version available (#5873)
Front Back
36. IBM
EXP 24S SAS Expansion Drawer (#5887)
– 2U disk drawer
– Up to 24 hot-swap SAS disks
– Configured as 1, 2 or 4 bays
– Choice of 146 GB, 300 GB and 600 GB drives
– High reliability as standard
– Redundant data paths to each slot
– Redundant hot-swap power/fan assemblies
– Redundant hot-swap Electronic Service Modules
– Attaches to SAS adapters and integrated controllers on selected POWER6
and POWER7 servers
37. IBM Power 795 (9119-FHB)
– Up to 256 POWER7 cores
– 24-192c @ 3.72 GHz (6-core SCM)
– 16-128c @ 4.25 GHz (TurboCore)
– 32-256c @ 4.0 GHz (MaxCore)
– New memory
– Up to 8 TB using quad-height dimms
– Enhanced RAS
– Memory Mirroring for Hypervisor
– Full thermal management capability
– Upgradeable from p6-595
– Serial number protection
– Reuse of frame, BPA, mid-plane, I/O
– New books, memory, system controllers
Power795
38. Hardware
Management
Console
Bulk Power Assembly and
24-port Ethernet hub
Central Electronics Complex (CEC)
24-256 cores, up to 8 TB Memory
Primary I/O drawer (required)
Optional I/O drawer,
batteries or media drawer
Optional I/O drawer
24-inch System Rack, 42U
Media drawer and lighting strip
Mid-plane
795 System Overview
39. SAP 2-Tier Performance
– Fewer processors to do the work
– Lower software costs
– Lower costs for power / cooling
– Lower maintenance costs
– Lower cost of acquisition
0
0.5
1
1.5
2
2.5
3
SAPs SAPs/core
128c POWER7 256c SPARC64 128c HP
At least 2.5x more performance per core!
40. Active Memory Mirroring for Hypervisor
– Eliminates system outages due to
memory dimm failures
– Maintains two identical copies of the
system hypervisor in memory at all times
– Both copies are simultaneously updated
with any changes
– In the event of a memory failure on the
primary copy, the second copy will be
automatically invoked and a notification
sent to IBM via the Electronic Service
Agent
– Available on Enterprise servers
40
Memory
41. The True Value of Large Systems
– Previous figures represented the
estimated performance of a single
AIX image using the entire system
– E.g. for a 64-core Power 795
– However, the aggregate
performance of a system running
multiple LPARs can be
significantly higher than this figure
indicates
LPAR
Size 16 32 48 64
64w 777
32w 434 868
16w 231 463 694 926
Total number of cores
Notes: These figures represent performance of dedicated partitions.
Starting with the introduction of the Power 795, a new rPerf estimate has been
added that represents multiple partitions of smaller sizes.
This new approach helps demonstrate the true value of large systems
E.g. the same 64-core machine
running four 16-core LPARs
42. Power 775 (9125-F2B/BT)
– Ultra-dense, water-cooled, packaging
– Up to 3,072 3.83 GHz cores per rack!
– 256 cores per node
– Up to 12 nodes per rack
– Up to 2TB memory per node
– Expansion
– 16 PCIe slots per node
– Integrated 10Gb Ethernet
– Integrated high-speed interconnect
– 1.9 TB/sec per node
– Performance
– ~80 TF per Rack
Power775supercomputer
48. HP sx2000 Oracle M9000 IBM Power 795
Enterprise Systems
IdeasInternationalRPE2
128-core
256-core
256-core
49. POWER7 SPARC64 VII+ T4 Itanium 9350
POWER7 HP SUN
RPE2 performance per Oracle license
Based on best result for each vendor (excludes Power TurboCore)
53. POWER7 SPARC64 T4 Opteron Nehalem-EX
Performance/kW (Best Available)
Energy Efficiency (IBM vs Oracle)
IBM Internal Use Only
Based on Ideas International for IBM Power 750, Sun/Oracle M3000, T4-2, X4640, X4470 11-Oct-11
54. Energy efficient computing
IBM Power 740
1x 16-core
10,512 kWh/yr
4 rack units
Oracle M5000
2x 32-core
47,304 kWh/yr
20 rack units
HP rx8640
2x 24-core
59,568 kWh/yr
34 rack units
Data from Ideas International
Configurations offering a similar level of performance
56. POWER7 EnergyScaleTM Functions
– Data collection and reporting
– Power usage data
– System temperature
– Standby state for inactive resources
– Processor nap mode
– Unassigned PCI slots
– Static Power Save
– Drop core frequency to pre-defined level
– Dynamic Power Saving modes
– Implemented by firmware, activated by Active
Energy Manager
– Partition level power management
– Allows LPARs to have different core
IBM Director Active
Energy Manager
Hardware
Management
Console
EnergyScale
Controller™
Flexible
Service
Processor
(FSP)
Power
Modules
POWER7
Core
Power Control Firmware
Power Module API
Power Mgmt Policies
Power
Measurement
57. Active Energy Manager
– Dynamic Power Saving (DPS)
– Adjusts clock speed in
accordance with system load
– Reduces energy consumption
across normal operating range
– Boosts performance when system
hits 100% utilisation
– DPS - Favouring Performance
– Boosts system performance at
the expense of a marginal
increase in energy consumption
– Dramatically reduces energy
consumption when system is
idle
Watts vs Throughput (32-core system)
500
600
700
800
900
1000
1100
1200
0% 20% 40% 60% 80%
Response Time vs Throughput
0
0.2
0.4
0.6
20% 40% 60% 80%
ftp://public.dhe.ibm.com/common/ssi/sa/wh/n/pow03042usen/POW03042USEN.PDF
Normal
With DPS
Normal
With DPS
58. Energy-efficient heat removal
Easy-to-mount rear door design that
attaches to customer-supplied water using
industry-standard fittings and couplings
Up to 17 kW (approximately 58,000
BTUs/hr) of heat removed from air exiting
the back of a fully populated rack
Optional for T42 racks and the 595
Mandatory for Power 575
Reduces heat to air by ~55%
Effectively doubles the kW/m2 for
customers constrained by air-conditioning
Rear Door Heat eXchanger
59. 21st Jan 1986
IBM RT Personal Computer Advanced Interactive eXecutive
60. – Strategic
– Committed roadmap
– Wide acceptance from key ISVs
– Reliable
– Dynamic configuration changes
– Help reduce planned and unplanned outages
– Integrated Logical Volume Manager + JFS2
– Easy to Manage
– Policies for allocation of server resource using Workload Partitions
– Create system images from pre-assembled templates with NIM
– Role-based Access Control and Integrated Systems Console
– High performance
– Scalability, automatic large page support, Asynchronous I/O, Direct I/O,
many others
Why AIX on POWER?
61. IBM AIX V7.1 highlights
Workload Partitions (WPARs)
Software-based virtualisation for workload isolation and management
Emulation of AIX V5.2 and V5.3 for older applications
Live Application Mobility to relocate WPARs between AIX images
Performance Enhancements
Active System Optimizer
Automatic async I/O tuning
Concurrent I/O FastPath
Encrypted JFS2 filesystem
Role-Based Access Control
Trusted AIX
Trusted Execution
“Secure by Default” installation
Kernel exploitation of Storage keys
Concurrent AIX kernel updates
Dynamic tracing with Probevue
Firmware assisted dumps
Live Dump
Support for >64-core LPARs
Terabyte segment support
Cluster Aware AIX
62. AIX 7 Active System Optimizer
WorkloadsSystem
Optimizer
AIX Kernel
Controls & Policies
PowerVM
Controls & Policies
HW
Controls & Policies
Monitor
Analyse
Modify
– Monitors AIX in real time via
continuous runtime analytics
– Ideal for systems with an often
changing workload
– Minimises use of system buses
– Increases cache hits
– Groups threads into affinity domains
– Runs like threads on the same core
– Reduces reference to remote
memory
– Migrates hot pages to local domains
– Standard part of AIX V7.1
– Available from TL01
– Process-level logging of ASO activity
63. AIX V7.1 Editions
– Express Edition
– Aimed at consolidating small workloads cost-effectively
– Can run alongside Standard Edition on the same machine
– Limited to 4 cores and 8 GB memory per core for each LPAR
– Limits are managed by IBM License Monitoring Tool (ILMT)
– Excludes Cluster Aware AIX
– Upgrade to Standard Edition is accomplished without a reboot
– Standard Edition
– Full function AIX, up to 256 cores and 8 TB memory for each LPAR
– Includes Cluster Aware AIX features and Profile Manager
– Enterprise Edition
– AIX Standard Edition +
64. AIX Workload Partitions (WPARs)
– Partitioned System Capacity
– Obtain a regulated share of the
processor and memory resources
– Share I/O devices, physical
processors, shared library and
text
– Architected for movement
between OS images
– System Partitions
– Separate administrative domains
with private logins, users, groups
– Shared or Dedicated /opt and /usr
– Can emulate older releases of AIX
– Application Partitions
– Encapsulation of a single
application
Workload
Partition
Application
Server
Workload
Partition
Web
Server
Workload
Partition
Billing
Workload
Partition
Test
Workload
Partition
BI
NetworksDisk or NFS storage
AIX
global Instance
Workload
Partition
Application
Server
Workload
Partition
Web
Server
Workload
Partition
Billing
Workload
Partition
Test
Workload
Partition
BI
NetworksDisk or NFS storage
AIX
global Instance
65. Reasons to use WPARs
– Save administrator effort
– Reduce the number of AIX instances
to patch
– Easy to monitor and control
– Lower memory requirements
– 68 MB vs 1GB for an LPAR
– Easy to deploy
– Takes seconds to create
– Rapid cloning possible
– Multi-system workload balancing
– Move WPARs between AIX global
instances using WPAR Manager
Workload
Partition
Application
Server
Workload
Partition
Web
Server
Workload
Partition
Billing
Workload
Partition
Test
Workload
Partition
BI
NetworksDisk or NFS storage
AIX
global Instance
Workload
Partition
Application
Server
Workload
Partition
Web
Server
Workload
Partition
Billing
Workload
Partition
Test
Workload
Partition
BI
NetworksDisk or NFS storage
AIX
global Instance
66. Versioned WPARs for AIX 7 (5765-H38/WP7)
– Simplify consolidation of older application environments
– Protects customer investment in application stacks
– Reclaim floor space and reduce hardware maintenance costs
– Get the benefit of POWER7 performance through SMT4
– Exploit features such as WPAR Mobility and Live Partition Mobility
– Allow older workloads access to newer storage/network technologies
– Limitations
– No support for NFS or HACMP within the WPAR
– Device support is limited to devices directly supported in WPAR (and on
POWER7)
– Offering includes limited fix support for AIX
– Requires concluding SP of final ML/TL
– Fix support is limited to security, data integrity problems and critical issues
and may not always be possible
–
67. AME
Physical
Memory
Effective
Memory
Without AME With AME
Expanded
Memory
Capacity
Active Memory Expansion
– Optional POWER7 server feature
– Compresses in-memory data
– Requires additional processor
capacity
– Amount varies according to workload and
desired compression ratio
– Configured on a per-LPAR basis
– Supports both dedicated and shared
memory LPARs
– Works with AIX V6.1 or V7.1
– Two possible uses
– Increase an LPAR’s effective memory
capacity for increased performance
– Reduce system memory requirement
to enable more LPARs per server
– Transparent to applications
AME
System Memory
LPAR LPAR
LPAR
LPAR LPAR
LPAR
LPAR
LPAR
LPAR
LPAR LPAR LPAR
System Memory
LPAR LPAR LPAR
LPAR
LPAR
LPAR
LPAR
LPAR
LPAR
LPAR
LPAR
LPAR
LPAR LPAR
69. PowerVM Editions: Features
Express Standard Enterprise
Management IVM HMC/SDMC HMC/SDMC
Virtual I/O Server
NPIV
Suspend/Resume
Shared Processor Pools
Thin Provisioning
Shared Storage Pools
Network Balancing
Active Memory Sharing +
Deduplication
Live Partition Mobility
NB: Active Memory Expansion is not part of PowerVM. It is available as an activation feature of POWER7 servers.
70. Shared Processor Partitioning
Micro-PartitionsTM
Share a pool of processors
Allocated processors “on demand”
Use no resource when idle
Get a guaranteed minimum
amount of resource when active
Can be capped or uncapped
Can be grouped together into
Virtual Shared pools
Can share disk adapters
Can share network adapters
Support dynamic reconfiguration
of memory
Are isolated and secure
EAL4+ certification and no
Common Vulnerability Exposures
I/OServer
I/O Adapters
LAN, WAN, …
6 cores
Disks
AIX6.1
AIX5.3
AIX6.1
Linux2.6
Hypervisor
Linux2.6
Pool 0 Pool 1 Pool 2
2
cores
AIX
71. Virtual I/O Server (VIOS)
AIX VIOSVIOS Linux
Transport Layer
(Hypervisor)
Allows sharing of adapters/devices
Virtual Ethernet
Virtual SCSI
Virtual Fibre Channel
Provides Active Memory Sharing
Pages on behalf of client partitions
Enables Live Partition Mobility
Provides memory copying services
Built on a version of AIX
Installed/managed as an LPAR
Dual VIOS configurations for
availability
Allows for rolling upgrades
IBM i
Virtual Adapters
Physical Adapters
and Devices
72. PowerVM Active Memory Sharing
Increases memory utilisation
Intelligently flows memory from one
partition to another according to load
Ideal for environments which become
active at different times of day
Significantly improves DR flexibility
Easy to control/monitor
Partitions have entitlements etc. similar to
CPU definitions
Commands/fields for identifying real and
virtual paging
Active Memory Deduplication
Detects and removes duplicate memory
pages to optimise memory usage
0
5
10
15
Europe Americas Asia
Time
MemoryUsage(GB)
0
5
10
15
Day Night
Time
MemoryUsage(GB)
73. Active Memory Expansion & Sharing
– Active Memory Expansion
– Effectively gives more memory
capacity to a partition using
compression technology
– Best fit when systems are memory
constrained
– Only available for AIX V6.1 or V7.1
partitions on POWER7 hardware
Active Memory Sharing
Moves memory from one VM to
another
Best fit when one partition is
not busy when another partition
is
Works for AIX, IBM i, and Linux
partitions
Active Memory Expansion Active Memory Sharing
Designed to allow more aggressive workload consolidation
AME Considerations
Value is dependent upon compressibility of data and available CPU
resource
Chargeable feature (free trial available)
Evaluation tool available in AIX V6.1 TL4
74. AIX I/O
Server
I/O
Server
LinuxIBM i
Virtual Ethernet
Inter-partition communication
Multiple virtual switches
Conforms to IEEE 802.1Q
Supports up to 4096 VLANs
Supports IPv4 and IPv6
Supports HMC and NIM
Access to external networks
via I/O Server
Shared Ethernet Adapter bridge
Link aggregation
Quality of Service controls
Mix virtual and real adapters
within a single LPAR
Hypervisor
75. Virtual SCSI
Virtual I/O Servers (VIOS)
Own real adapters
SAS, FC, FCoE, SCSI, iSCSI, USB
Support disks, optical media and
internal tape drives
Support multi-pathing
Connect to LPARs via virtual
(VSCSI) adapters
Logical Devices
Appear to client as SCSI devices
Virtual disks
Can be used as boot devices
Support mirroring and multi-pathing
(MPIO) for availability
Can be shared between LPARs
A A’
A’A
AIX I/O
Server
I/O
Server
IBM i Linux
Database
Hypervisor
76. IBM Confidential
PowerVM Storage Pools
– Extending the virtualisation layer beyond a single system
– Combines vSCSI with clustering technology & a distributed data object
repository
Hypervisor Hypervisor Hypervisor
Hypervisor Hypervisor Hypervisor
VIOS
Storage Pool of SAN & NAS
VIOS VIOS VIOS VIOS VIOS
VIOS VIOSVIOS VIOSVIOS VIOS
77. Thin provisioning
The amount of physical storage used is less than the amount
defined for the virtual workload,
Optimal storage utilisation across shared storage pool.
Additional storage is provided dynamically when workloads expand
Storage is released when workloads contract.
Automates optimised storage utilisation
More cost-efficient use of storage resources
Integrates multiple storage subsystems
78. Virtual Fibre Channel
– Virtual I/O Servers (VIOS)
– Own physical FC HBAs
– Must be NPIV capable
– Create virtual NPIV adapters
– Pairs with unique WWPNs
– Create connections to physical
HBAs using FC Passthru
– Client LPARs
– Have direct visibility of SAN
zoning and LUN masking
– Can use MPIO
– Active-active load balancing
– Mixed dedicated and virtual paths for
Live Partition Mobility
AIX I/O
Server
I/O
Server
Linux
Hypervisor
79. Integrated Virtualization Manager
– Create and manage Micro-partitions
using a browser interface
– Available on Power Blades and the Power
710/730, 720, 740 and 750
– Intuitive, user-friendly interface
– Packaged with VIOS
– Subset of HMC functionality
– Single VIOS
– No support for concurrent firmware
maintenance
– Can be migrated to an HMC environment
at a later date
80. Hardware Management Console
– Manages Advanced Functions
– Required for Enterprise servers
– Optional for express servers except blades
– Needed for Dynamic LPAR, On/Off CoD, concurrent
maintenance and complex environments
– Provides remote power management
– Connects to Service Processor using Ethernet
– Provides virtual console for LPARs
– Using virtual serial connections
– Consolidates error logs
– Connects to LPARs by real or virtual Ethernet
– Support for up to 32 servers or 254 LPARs
IBM
H C R U6
IBM
System
p5
IBM
System
p5
IBM
System
p5
IBM
H C R U6
IBM
H C R U6
IBM
System
p5
IBM
System
p5
IBM
System
p5
server
82. 82
IBM Capacity on Demand offerings
Permanent
For customers who foresee growth
and want hardware resources
installed for easy activation over
time
Temporary
For customers with predictable
bursts in utilisation (e.g. end of the
month or year) who want to pay for
additional capacity on a daily basis
Utility
For customers who experience
unpredictable short bursts in
utilisation and who want to pay for
additional capacity by the minute
85. Fault avoidance - designed to avoid downtime
Diagnose failures as they happen
Resolve intermittent faults with no
intervention
De-allocate system
resources if necessary
Utilise system
redundancies
Keep the System Operational
86. Dramatic reduction in
unplanned outages
Self-correcting
hardware
Concurrent
maintenance of
hardware and
firmware
Complete isolation of
Logical Partitions
Resilient and Highly Available Power Systems
Over 40 years
of IBM large
system
innovation
Kernel Storage Keys
Concurrent AIX updates
Functional Recovery Routines
Workload Partitions + Application
mobility
+
POWER6/7
Systems Alternate Processor
Recovery
Hypervisor mirroring
Live Partition
Mobility
Hot add, repair and
maintenance of
books
POWER4/5
Systems
Power780
87. First Failure Data Capture
– Massive investment in design
– Error checkers provide real-time
machine status information
– Data is continually monitored by
Service Processor
– Can phone IBM if repair needed
– Many errors can be healed or
functions varied off while the
server remains available
– Suspect components
automatically de-configured
during reboot following failure
Service
Processor
Non-volatile
RAM
CPU
L1
L2/L3
Memory
Disk
Log error
Error checkers
88. Memory scrubbing for soft single
bit errors that are corrected in the
background while memory is idle,
to help prevent multiple bit errors.
Dynamically reassign memory
I/O via bit steering if error
threshold is reached on same bit.
Failing memory bit steered
to spare memory chip
Chipkill
Bit scattering allows normal single bit ECC error processing, thereby keeping the
system running with a chipkill failure.
Bit steering allows memory lines from a spare memory chip to be dynamically
reassigned to a faulty line in a memory module to keep the system running.
If all bits are used up on the spare memory chips, and the threshold is reached, the
Service Processor will be invoked to request deferred maintenance.
Scatter memory chip bits across four separate ECC words for chipkill recovery
Spare
Memory
chipX X
Chipkill Memory
89. Dynamic Processor Sparing
Running on 'n' processors
Keeps running on 'n-1'
Back to running on 'n'
With CUoD
Hypervisor assigns
inactive CUoD processor
to get System back to full
capacity
Without CUoD
Processor is dynamically
de-configured and
partition configurations are
automatically adjusted.
Applications continue to
run across a processor
failure
Hard processor failure or
dynamic de-configuration
(preventative action)
90. POWER7 Processor Recovery
– Processor Instruction Retry
– Allows the processor core to
recover completely from transient
faults
– Alternate Processor Recovery
– For solid faults, moves the
instruction stream to another core
– Partition Availability Priority
– If no spare core exists then low-
priority partitions can be reduced in
size or terminated in order to
maintain high-priority services
256KB L2
IFU
CRU/BRU
ISU +
integrated
Recovery
Unit
DFU
FXU
VSX
FPU
LSU
91. Health Care for a Power System Server
Service Processor
Electronic Service Agent
CE Dispatch
System p
IBM
Processors
Memory
Power
Thermal
Blowers/Fans
Central
repository for
all system
errors
On-board in the CEC.
Runs asynchronously.
Does not degrade CEC
performance.
AIX device drivers
Provide error
Information to the
AIX error log
Phones latest error
Information to
CE/customer when
necessary
Initiates maintenance
Action with the parts
Required to fix the
real problem
Cycles continuously
looking for faults
Input / Output
CEC
HMC Service
Focal Point
AIX Error Log
92. Server Downtime
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
Hoursperyear
Medium Large V. Large
Organisation Size
Downtime per server
AIX Linux
0
5
10
15
20
25
30
35
40
Hoursperyear
Medium Large V. Large
Organisation Size
Total Downtime
AIX Linux
Adapted from “Does your OS Matter?” - Solitaire Interglobal Ltd. October 2011
93. ► All memory is accessible by all OS
code (kernel + device drivers)
► Difficult to debug memory over-
writes
POWER5
► Multiple domains provide isolation
among different AIX components
► Facilities also provided for
middleware / applications
POWER6/7
OS & Middleware Component Isolation
Kernel
JFS
VMM
LVM etc.
IBM DDs
Ethernet
FC-AL
SCSI etc.
ISV DDs
Veritas
PowerPath
Etc.
AIX V6.1
Kernel
JFS
VMM
LVM etc.
IBM DDs
Ethernet
FC-AL
SCSI etc.
ISV DDs
Veritas
PowerPath
Etc.
AIX V5.3
DB2
Components
Application
Code
Middleware e.g. DB2
Other
Middleware
DB2
Components
Application
Code
Middleware e.g. DB2
Other
Middleware
65% of AIX V5 crashes were caused by memory over-writes
Memory Protect Keys
94. Security breaches - Economic Impact
0
500
1000
1500
2000
2500
3000
3500
Thousands(USD)
Medium Large V. Large
Organisation Size
Average Annual Cost (reported)
AIX Linux
Adapted from “Does your OS Matter?” - Solitaire Interglobal Ltd. October 2011
– The cost of a security breach
– Strengthening existing IT security
and carrying out additional
training
– Contacting those whose records
may have been exposed
– Credit monitoring for those
affected
– Legal action taken by people who
may have suffered a financial loss
– Damage to the company/brand
reputation
– Email blacklisting
– Impact on share price
– Costs to regain market position
David Hobson, managing director of Global Secure Systems – SC Magazing
95. Power and AIX Standard Security
– PowerVM Hypervisor
– Part of our digitally signed firmware with strong cryptography which makes
it impossible to remotely install a modified fileset into the EPROMs
– AIX
– Certified to meet EAL4+ Common Criteria
– “Secure by Default” installation option
– Trusted Execution insures that only programs signed by trusted entities can
run; prevents even the root user from modifying the system
– Encrypted File System protects data even from a malicious root user
– Role Based Access Control eliminates the need for root privileges
– Stack Execution Disable prevents stack based buffer overflow attacks
– Cryptographic Services provide simplified, secure management of
encryption keys and digital certificates
96. PowerSC (5765-G82/PSE)
– Express Edition
– System Compliance Profiles
– Ensures that the settings in the OS match security standards such as PCIDSS
and COBIT
– Standard Edition
– Trusted Boot
– Ensures that an OS running on a Power System has not been inadvertently or
maliciously altered to compromise the security of the system
– Trusted Logging
– Provides a central tamper proof repository for the system logs for the all the
virtual machines running on a Power system
– Trusted Network Connect
– Ensures that during the virtual system network connection every virtual system
is at a site-specified security patch level and update level
97. LPAR-1
Shared Pool.
LPAR-3 LPAR-4LPAR-1 LPAR-2 LPAR-3 LPAR-4
AIX Kernel AIX Kernel AIX Kernel AIX Kernel
V
I
O
S
V
I
O
S
Pro 1
Pro 3
Pro 4
SAN
Shared Pool
Gb Ethernet
P P P P P P P P
P P
P P PP P P P
P P P
Data
P P
P P
P
Pre-requisites
POWER6/7 hardware with
PowerVM Enterprise Edition
Full virtualisation
SAN boot
Shared network and SAN
Partition profile is copied
and partition is created
AIX Kernel AIX Kernel AIX Kernel AIX Kernel
LPAR-2
App A
P P
P P
P
App A
Memory is copied
asynchronously until
threshold reached
Application is switched to
target server
Remaining pages are
copied synchronously
Small performance impact
Source partition is deletedAny size partition may be
moved
Time taken is ~3GB/minute
using Gb Ethernet
Boot
Pro 2Pro 2
Live Partition Mobility
98. Suspend/resume simplifies administration
– How it works:
– “Freezes” a VM and saves the complete system state
– After suspension, the server resources are freed up for use by other
workloads
– The entire VM system state is stored in a set of files and can be resumed on
the same server or on a different system after migration exactly where it left
off, without data loss
– Client benefits:
– Resource balancing –suspend low-priority or long-running workloads to
allow more urgent processes access to server resources temporarily
– Simplified maintenance –administrators can perform system updates or
CEC upgrades without having to spend time on workload shutdown/startup
processes
– Debug/forensics –a workload can be temporarily suspended and a copy
made for offline analysis for security or performance purposes
99. Simplifies continuous application availability
Manages planned & unplanned outages
Standard Edition
Data Centre solution (up to 40 Km)
Shared storage clustering
Automatically reacts to events
Multiple heartbeats
Supports both active-active and active-standby
No configuration restrictions
Enterprise Edition
Multi-site solution (unlimited distance)
Adds sync/async data mirroring
AIX GLVM (IP network)
DS8000 Metro-Mirror (PPRC)
SVC Global Mirror
EMC SRDF / Hitachi Truecopy
Power780
Power780
Power780
Power780
PowerHA SystemMirror for AIX
100. Ideal for existing Symantec
customers migrating from SUN to
Power
http://go.symantec.com/ibm
• Products: Veritas Storage Foundation
& Veritas Cluster Server
• OS: AIX, Red Hat and Novell SUSE
Symantec Veritas
Storage Foundation HA
for Linux on PowerTivoli System Automation for
Multiplatforms (SA MP)
Availability Options for Linux on Power
http://www.ibm.com/software/tivoli/products/sys-auto-multi/
• OS: AIX, Red Hat and Novell SUSE
IBM Tivoli developed High
Availability and Automation solution
for mission-critical servers and
workloads
101. PowerHA PureScale for AIX
– Based on mainframe Sysplex
– Provides structured data models and
distributed services
– Global Lock Manager
– Low-latency Infiniband interconnects
– Based on RDMA and uADPL sockets
– Runs in an LPAR or dedicated server
– Supported on all POWER7 servers
– Used with DB2 and GPFS provides
– Continuous database availability
– Dynamically scalable database capacity
– DB2 pay-as-you-use charging model
– A cost-effective replacement for RAC
“from scalability to flexibility, through ease of use and high availability, to cost ... IBM appears to offer significant advantages”
IBM pureScale Application System vs Oracle Exadata X2-2 - Bloor Research, Dec 2010
102. General Parallel File System
– High-performance, scalable filesystem
– Files span multiple nodes, multiple disks
– High aggregate bandwidth
– Eliminates all single points of failure
– No requirement for PowerHA
– File access from any node
– Network access via NFS
– Many commercial applications
– Digital media
– Business intelligence
– Financial analysis
– Available for AIX, Linux and Windows
103. Systems Director
Express Edition
Systems Director
Enterprise Edition
App
OS
Image
App
OS
Image
Systems Director
Standard Edition
IBM Systems Director Editions
Analyse and report
historical performance
Perform capacity
planning
Automate configuration
and placement for new
workloads
Manage workload
availability end-to-end
Monitor system health
Provide threshold and
error alerts
Update OS’s and
firmware
Visualise
physical/virtual system
relationships
Understand and
control energy use
within existing
capacity
Simplify deployment
with virtual images
Monitor network
system health with
servers and storage
+
+
104. Systems Director Express Edition
– Systems Director base
– Discovers virtual and physical systems and collects inventory data
– Visualises relationships to other resources in the network
– Monitors hardware, power, and compliance status for systems
– Provides simplified tasks to download and install updates
– Automatically performs custom actions in response to events
– Provides a single framework for configuring managed resources
– Provides a set of integrated tools that support remote access
– VMControl Express
– Creates, modifies and deletes logical partitions
– Relocates logical partitions using Live Partition Mobility
– Service and Support Manager
105. Systems Director Standard Edition
– Active Energy Manager
– Monitors and manages energy usage at the system, chassis or rack level
– Displays energy use for systems, or groups of systems, over time
– Displays information on the inlet and exhaust temperatures for systems
– Allows a client to set a maximum energy level per system (capping)
– Implements Static and Dynamic power saving modes
– VMControl Standard Edition
– Creates virtual images that encapsulate OS, middleware, and applications
– Maintains virtual images in a centralised library
– Deploys virtual images instead of separate software components
– Network Control*
– Discovers and collects inventory data for NICs and switches *limited functionality
App
OS
Image
App
OS
Image
106. Systems Director Enterprise Edition
– VMControl Enterprise Edition
– Manages groups of virtual appliances across multiple physical servers
– Provides a dashboard showing health and status of deployed workloads
– Automates virtual image mobility for optimal utilisation and resilience
– Tivoli Monitoring*
– Provides mapping of storage and network resources to client LPARs
– Provides storage and networks utilisation statistics for the VIOS
– Provides comprehensive performance statistics for LPARs and the frame
– Provides predictive trending to manage performance over time
– Monitors energy metrics, provides alerts and takes action in response
– Tivoli Application Discovery and Dependency Manager*
– Determines if configurations comply with your policies *limited license applies
107. Virtualisation of IT as
a service - “cloud”
Business-driven
service
management
Service oriented
delivery of IT
Maximising the value of a Shared infrastructure
Physical consolidation
Virtualisation of systems
Systems, network and
energy management
Highly virtualised
resource pools for
mobility
Integrated IT service
management
Green by design
108. Where everything in the Data Centre becomes
Virtualised - driving higher utilisation and lower costs
Reusable, shared and optimised
Resilient, secure and highly available
Scaleable and responsive
Flexible enough to support rapidly changing business requirements
Managed by business priorities
Hypervisors Network
Intel
Storage
POWER
Private Clouds
Tivoli + Director + PowerVM: technology to meet today’s challenges
109. Industrial strength
virtualisation coupled
with automated resource
balancing and virtual
image management
Integrated service
management
platform with
automated IT service
deployment, full
lifecycle
management,
metering &
chargeback
Basic cloud functions
including simple self
service interface and
infrastructure with
automated provisioning
Cloud Capabilities
DeliverITwithoutBoundaries
Virtualization
Foundation
Entry Cloud
Advanced Cloud
Cloudburst on Power & IBM
Service Delivery Manager
POWER7 systems, PowerVM,
PowerSC, Systems Director &
VMControl
IBM SmartCloud Entry
delivered by IBM Starter Kit for Cloud
Infrastructure as a Service Technologies
Infrastructure
Platform
Usage and
Accounting
Availability and
Performance
Management
and Administration
Security and
Compliance
Power Systems Cloud Solutions
111. Why Linux on POWER?
– Save ~$30,000 per 100 users per year*
– Reduce infrastructure costs
– Includes hardware, software, space, power and cooling
– Increase IT staff productivity
– Savings in data centre operations, sys admin and help desk
– Increase end-user productivity
– Reduce application outages by a factor of 10
– Pay-back period 6.3 months
– Includes hardware, software, training and consultancy costs
– A full range of applications
– Oracle, Sybase, DB2 …
– Tivoli, MQSeries, Websphere …
– IBM Service and Support
*Adding Business Value with Cross-Platform Solutions: Linux Running on IBM Servers – IDC, October 2008
112. PowerVM features supported on Linux
Feature 10 SP3 11 SP1 5.6 6
Micro-partitions Yes Yes Yes Yes
Dynamic LPAR Processors Yes Yes Yes Yes
Dynamic SMT enable/disable Yes Yes Yes Yes
Dynamic LPAR Memory Add only Yes Add only Yes
Dynamic LPAR I/O Yes Yes Yes Yes
Virtual Ethernet & SCSI Yes Yes Yes Yes
Virtual LAN Yes Yes Yes Yes
IBM i Hosted Virtual I/O Yes Yes Yes Yes
Active Memory Sharing No Yes No Yes
Active Memory Expansion No 2012 No No
NPIV Yes Yes Yes Yes
Virtual Tape Yes Yes Yes Yes
Live Partition Mobility Yes Yes Yes Yes
POWER7 VSX No Yes No Yes
113. POWER6 Withdrawals Summary
Product Name Machine / Model Effective withdrawal date
12X Drawer 7314-G30 28-Aug-2009
Power 560 8234-EMA 7-Jan-2011
Power 570 9117-MMA 7-Jan-2011
JS23/JS43 7778-23X 29-Apr-2011
Power 520
Power 550
8203-E4A
8203-E8A
27-May-2011
Power 595 9119-FHA 29-Jul-2011
Power 560 8234-EMA 7-Jan-2011
Power 570 9117-MMA 7-Jan-2011
Power 575 9125-F2A 30-Jul-2010
JS23/JS43 7778-23X 29-Apr-2011
114. Link to IBM Systems Information Centre
http://publib.boulder.ibm.com/infocenter/systems/scope/hw/index.jsp
Link to AIX Operating System Certifications
http://www-03.ibm.com/systems/power/software/aix/certifications/index.html
Link to Power Systems Service and Support Best Practices
http://www14.software.ibm.com/webapp/set2/sas/f/best/home.html
Link to AIX Operating System Service Strategy Details and Best Practices Guide
http://www14.software.ibm.com/webapp/set2/sas/f/best/home.html
Link to IBM Software Support Handbook
http://www14.software.ibm.com/webapp/set2/sas/f/handbook/prevention.html
Link to Product Lifecycle Management website
http://www-01.ibm.com/software/support/lifecycle/index.html
Link to Fix for planned service end dates for AIX V5.3 Technology Levels
http://www-933.ibm.com/eserver/support/fixes/fixcentral/pfixpacks/53
Link to Fix for planned service end dates for AIX V6.1 Technology Levels
http://www-933.ibm.com/eserver/support/fixes/fixcentral/pfixpacks/61
Link to Fix for planned service end dates for AIX V7.1 Technology Levels
http://www-933.ibm.com/eserver/support/fixes/fixcentral/pfixpacks/71
Useful links
115. Summary
– IBM continues to deliver the most complete and flexible Unix platform for
the enterprise
– Choice of AIX, Linux and IBM i
– Massive scalability
– Industry leading virtualisation, built-in
– Pay-as-you use and pay-as-you-grow options
– Industry leading Reliability, Availability and Serviceability
116. Sample Questions
– Which of the following describes a benefit of TurboCore?
– A. Increases the system bandwidth
– B. Increases the speed of the core
– C. Increases L2 cache available per chip
– D. Increases the threads per core from 4 to 8
– Answer: B
116
117. A customer has a 2-core POWER5 server, thirty x86-based Linux servers,
and 10TB of storage. They have budget approval for a complete
infrastructure refresh, including a disaster recovery system.
Which solution supports the customer needs at lowest cost?
– A. Blade servers with DS5000
– B. Power 770 servers with DS8000
– C. Power 730 servers with DS5000
– D. Power 755 servers with DS6000
– Answer C
117
118. The IT manager has asked for assistance in building a proposal to
consolidate their Sun servers running WebSphere to an existing Power
570. The CFO has received a proposal for an HP blade
– solution to replace the Sun servers. What is the correct approach?
– A. Request a meeting with the IT manager and CFO
– B. Provide IBM customer references and white papers
– C. Propose a POWER7 solution with multiple partitions
– D. Provide IBM Global Services and financing for the HP solution
118
Answer: A
119. The CIO has recommended that their system be upgraded to a more
current Power Systems solution. The existing lease has 9 months before
it ends. What is the least expensive financing solution for the client?
– A. Add the upgrade to the remaining lease term.
– B. Buy out the existing lease and finance the upgrade.
– C. Buy out the existing lease and finance a new system.
– D. Provide a 3 year base lease extension and finance the upgrade.
119
Answer: D
120. A customer is running an Oracle database on a 16-core POWER5 570 and
wants to migrate to a Power 750. Processing requirements have
remained stable. How can the sales representative help to justify the
purchase?
– A. Active Energy Manager found in the Power 750 firmware lowers
energy cost.
– B. The Power 750 requires fewer cores to run the Oracle database,
lowering software costs.
– C. The ability to concurrently replace cores and memory on the
Power 750 reduces the cost of disrupted service.
– D. The Power 750 has better virtualization features than the
POWER5 570, increasing the utilization of the system.
120
Answer: C
