|
IBM United States
Hardware Announcement 101-285 October 4, 2001
IBM
|
pSeries 690 Brings POWER4 Technology to the 32-Way
Server Environment
 ENUS101-285.PDF (149KB)
|
At a Glance
IBM Innovative Technology — A top performing UNIX server featuring:
- IBM SOI/copper chip technology
- Mainframe-inspired LPAR and self-managing capabilities
Cost-efficient growth path to the future
-
64-bit system scalability
- 8-, 16-, 24-, or 32-way — packaged on four Multi-Chip Modules (four or eight POWER4 processors per MCM)
- 1.1 or 1.3 GHz POWER4 processors
- 128 MB of L3 ECC cache per MCM
- DDR System memory expandable from 8 GB to 256 GB
- Up to eight 7040-61D I/O drawers per server; and supporting 20 blind-swap PCI bus slots and up to 16 Ultra3 SCSI disks that can be hot-plugged
- Hardware split into as many as 16 LPARs, each functioning as a "computer within a computer" with its instance of the operating system
For ordering, contact:
Your IBM representative, an IBM Business Partner, or IBM Americas Call Centers at 800-IBM-CALL (Reference: RE001).
Overview
Innovative IBM 
platforms with new tools and application flexibility
help lower costs, improve efficiency, and expedite e-business
transformation.
- Innovative Technology — Mainframe-inspired features extend IBM's leadership in open UNIX® systems
-
New Tools
— e-business infrastructure with self-managing IBM
platforms
- Application Flexibility — Uncompromised flexibility in selecting, building, and deploying e-business infrastructure applications
The IBM pSeries 690 (7040-681) is the first in a family of
products representing a new generation of IBM's UNIX-based, 64-bit,
symmetric multiprocessing (SMP) servers. It gives you the power you need
to consolidate workloads within a single, powerful, UNIX-based system.
The large SMP server no longer functions merely as a server, but rather
as self-contained "IT infrastructure in a box."
The attributes of the pSeries 690 set the stage to grow as your needs change:
- High availability
- Scalability
- Reliability
- Flexibility
Use it as you make your transition from existing RS/6000® Model S80 and p680 units or move to a brand new computing model.
New platform designs and new features (for example, world-class SMP performance, serviceability, assurances, and logical partitioning), position the pSeries 690 to satisfy the unpredictable and unprecedented performance, capacity, and availability you demand. It can also handle high-growth e-business, NetGen segments, the Telco Customer Care and Billing (CCB) subsegment of CRM, and the high performance computing environment.
Key benefits:
-
Fewer systems to manage and maintain
- Less floor space and power
- Reduced likelihood of operator error
-
Greater flexibility in:
- Using available capacity
- Matching dynamically changing workloads
- Managing rolling software upgrades
- High availability by repartitioning your system
- Enhanced management using of the Hardware Management Console
- Hardware and software investment protection in supporting multiple instances of the AIX® operating system and applications
You gain leading-edge technologies, such as the POWER4 microprocessor, logical partitioning, and enhanced management features to bolster continuous operation and workload management in your 24 x 7 business environment.
Planned Availability Date
December 14, 2001
April 25, 2002
- 24-way, 1.1 GHz configurations
- 24-way, 1.3 GHz configurations
- pSeries 690 configurations with 7 or 8 7040-61D I/O Drawers
March 29, 2002
- HACMP Version 4.4.1 Support
- Processor Feature Conversions
Key Prerequisites
AIX 5L for POWER V5.1 with the 5100-01 Recommended Maintenance package (APAR IY21957), or later
Description
Innovative Technology
IBM is delivering the most technologically advanced line
of servers in the industry. Each of the servers in the family provide
outstanding performance and scalability resulting in industry-leading
performance and price/performance across a wide range of e-business and
technical computing workloads.
Today's announcements further extend IBM's leadership with enhanced processors, self-managing, LPAR and configuration capabilities for its pSeries of UNIX systems.
-
Designed for the self-managed environment with high levels of
reliability, availability, and serviceability (RAS) capability
- Self-configuring — Hot-plug disk and PCI
- Self-healing — Chipkill™, memory bit scrubbing, dynamic deallocation of processors, cache, PCI bus, and LPAR
- Self-optimizing — Workload Manager, LPAR
- Self-protecting — Cryptographic coprocessors, digital certificates, Kerberos
-
Leadership technology
- Copper and silicon-on-insulator (SOI) helps produce faster chips that run cooler
- POWER4 is the first SMP on a chip with dual microprocessors
- POWER4 is the fastest microprocessor based on SPECcpu2000 benchmark results
- Up to 1.3 GHz, 64-bit processors
-
Flexible configurations
- SMP, LPAR, clusters
The pSeries 690 will be able to attach to SP™ systems over time. The design of the Central Electronics Complex (CEC) is intended to eliminate the concept of unique nodes. Future SP implementations are intended to be configurations of clustered pSeries 690 servers. For the first time, SP CECs and base server designs will be indistinguishable. The combination of LPAR, common CEC hardware, platform-optimized I/O drawers, a common hardware console, common switch fabric, and attachment adapters will provide customers with new levels of flexibility and scalability not previously available in the market.
For additional information on SP and PSSP cluster configuration support on pSeries 690 servers, refer to Hardware Announcement 101-283 , dated October 4, 2001.
IBM anticipates that one or more Linux distributors will support 64-bit Linux running in an LPAR by first quarter 2002. The support will allow 32-bit applications to run in partitions with a maximum of four processors. For more information about Linux running on pSeries systems, consult:
Linux for pSeries is available directly from the Linux distributors. IBM
does not sell nor install Linux on pSeries systems.
pSeries
The pSeries is the most powerful technologically advanced family of UNIX servers. pSeries meets a diverse set of needs scaling from small departmental and distributed systems to massively parallel supercomputers. Each pSeries features IBM's UNIX operating system AIX 5L, the most open UNIX with Linux Affinity.
The ultra-powerful pSeries 690 sets the standard for datacenter-class UNIX servers with mainframe-heritage LPAR and self-managing features.
The pSeries 690 is the first system to offer the POWER4 microprocessor. The POWER4 chip is the first ever dual processor SMP on a single piece of silicon. It owes its industry-leading performance to IBM's advanced SOI and copper fabrication technology. The balanced performance of the POWER4 chip makes it equally at home with mission-critical commercial or compute-intensive applications.
The system features include:
- Full 64-bit implementation
- AIX 5L Version 5.1 and HACMP support
- POWER4 processors operating at 1.1 GHz or 1.3 GHz
- 8-, 16-, 24-, 32-way SMPs
- 8 GB to 256 GB DDR memory
- L2 and L3 ECC caches
- Up to 16 logical partitions (LPAR)
- Service processor for RAS
- First Failure Data Capture
- Self-diagnosing and healing features
- IBM Hardware Management Console for pSeries (HMC)
The pSeries 690 can be configured up to eight 7040-61D I/O drawers. The I/O drawers can contain up to 16 hot-swap capable disks structured into four 4-packs of 18.2 or 36.4 GB disks. This delivers up to 582 MB maximum per I/O drawer, with up to eight drawers and 4.65 terabytes of internal disk for the system.
The drawer contains two planars with 10 slots per planar, 20 slots/drawer, with a maximum of 160 slots per system. The slots are hot-pluggable to permit PCI adapters to be added or replaced without extending the I/O drawer, all while the system remains available to the customer.
For unmatched UNIX system performance and scalability, up to 16 POWER4 chips (32-way) can be added to the pSeries 690. To help customers consolidate UNIX and Linux workloads the system can be split into as many as 16 logical partitions.
Complementing the power of the pSeries 690 is extraordinary self-managing
capabilities. The system can detect faulty memory, processors, cache,
and PCI buses and dynamically take them offline. It even makes the
service call. All of this without requiring administrative action or
interrupting operations. Growth is easy with hot-swappable disk drive
and hot-plug PCI. For the ultimate in availability the systems can be
clustered together with High-Availability Cluster Multiprocessing (HACMP)
software, the leader in UNIX availability solutions.
Self-Managing Systems
Complexity is a way of life for businesses today. On average, companies must contend with five different architectures in their e-business infrastructure. The challenge customers face is how to manage all of these resources cost effectively while transforming their infrastructure. Compounding the problem is the projected shortfall of skills and the explosive growth of the Internet.
This may sound like a daunting task, but IBM is designing its IBM
systems to be self-managing so customers can concentrate on their
business needs. Project eLisa is IBM's blueprint for self-managing
systems. The goal is to use technology to manage technology, creating an
intelligent, self-managing IT infrastructure that minimizes complexity.
This will enable you to manage environments hundreds of times more
complex and more broadly distributed than exist today.
Project eLisa-based systems will allow you to grow your computing resources without increasing critical staff and costs to manage larger more complex computer networks.
This all-encompassing strategy addresses the self-management needs in four areas:
- Self-configuring — systems can be reconfigured or expanded without interrupting operations
- Self-healing — systems can detect impending problems and correct for them automatically to keep systems going and lower administrative expense
- Self-optimizing — systems can dynamically balance resources and workloads to maximize responsiveness and throughput
- Self-protecting — systems can help protect transactions and enterprise data
I/O Drawer
The 7040-61D I/O drawer provides PCI adapter slots and internal disk capabilities for use with pSeries servers such as the pSeries 690 (7040-681). The Model 61D is a 4U drawer that mounts in the 24-inch 7040-61R System Rack and connects to the system Central Electronics Complex via remote I/O cables.
Each Model 61D I/O drawer provides twenty blind-swap PCI slots and
sixteen hot-swap disk bays. Redundant power converters and power cabling
ensure high reliability and availability. It utilizes 350 V bulk power
supplied from the 7040-61R system rack.
System Rack
The 7040-61R System Rack provides space for mounting system drawer components for pSeries servers such as the pSeries 690 Central Electronics Complex (7040-681) and the 7040-61D I/O drawers. The Model 61R is a 24-inch rack with 42U of rack space.
The 7040-61R utilizes a 350 V dc bulk power subsystem to provide power for the components within the rack. The bulk power subsystem incorporates redundant bulk power assemblies mounted in the front and rear sections of the top 8U of the rack.
If additional 24-inch rack space is required an optional expansion rack
feature is available for the Model 61R rack. The expansion rack
feature connects to the side of the primary rack to provide an additional
42U of rack space. The components in the expansion rack are powered from
the bulk power subsystem in the primary Model 61R rack.
RAS Features and Functions
Excellent quality and reliability features are the keys to all facets of pSeries products. The measures taken help ensure the products operate when required, perform reliably, efficiently handle infrequent failures in a nondisruptive fashion, and provide timely and competent repair to allow resumption of operations with minimal inconvenience.
Reliability begins with the development of architectures, where RAS innovations are of paramount importance. The reliability concept flows through design and product development stages, where RAS designs are reviewed, assessed, developed, evaluated, and optimized. It continues through the manufacturing and release processes, where the manufacturing quality is extensively measured and is under continual evaluation. It culminates in service and support, where:
- Reliability is consistently monitored for deviation from the criteria
- Warranty and maintenance have high priority
- Significant customer problems are assigned to and addressed by an expert team
pSeries 690 system processes, from the architectural and concept phases
of development through the provision of service and support, are ISO
certified and audited periodically for ISO compliance by representatives
of Underwriters Laboratories.
Reliability — Fault Avoidance
Major effort has been spent on the development of the pSeries 690 system to analyze single points-of-failure within the CEC (comprised of the processors, memory, and the I/O interface) and to either eliminate them or to provide hardening capabilities to significantly reduce the probability of failure. The best way to harden a system is to prevent errors from occurring. Components within the CEC were selected to provide "server" levels of reliability. These components undergo the highest level of reliability testing within IBM. During the design and development process, subsystems go through rigorous verification and integration testing processes. During system manufacturing, systems again go through a thorough testing process to ensure highest product quality level.
Extensive error detecting and checking circuitry is included to help ensure that infrequent component failures are recognized and corrected automatically, and properly reported for isolation and repair. This helps preserve the integrity of data stored and transported in the system. To support a diagnostic technique called "run time first failure data capture," errors detected by the CEC error checking circuitry are automatically captured and analyzed by the service processor at the time of error. The system is designed to identify and log the appropriate replacement part or parts without reproducing the error condition. This can significantly reduce the repeat failure and repair scenarios associated with error recreate strategies by providing the correct isolation at the time of the failure. Reduced outages and quicker repair time can result in higher system reliability and better overall availability.
Hardware and software failures are recorded in the AIX log facility and
analyzed by error log analysis (ELA) routines to provide warning to the
system administrator on the causes of system problems. On pSeries 690
systems, all service action requests are forwarded from the operating
system partitions to a Service Focal Point (SFP) application running on
the Hardware Management Console (HMC). The service focal point
application analyzes the potentially duplicate reported failures from the
various logical partitions, attempts to determine the root cause of the
fault, and upon identifying the cause, initiates call-home requests to
IBM support via the HMC modem and the Service Agent application program.
This enables IBM service representatives to have replacement hardware
components available when a service call is placed, minimizing system
repair time.
Availability — Backup and Replication
Power and Cooling Subsystem: The pSeries 690 brings new a level of availability features and functions to the Enterprise Server. Within the CEC Rack, the N+1 power and cooling subsystem can provide complete redundancy in case of failures in the bulk or regulated power supplies, the power controllers and the cooling units. Concurrent repair is supported on all of the power and cooling components.
The pSeries 690 also features redundant Internal Battery Features (IBF) to hold the system up during brown-out conditions. For full power loss protection, the pSeries 690 supports optional uninterruptible power supply (UPS) systems in addition to, or in place of, the IBF features.
In the I/O drawers, N+1 power and cooling is provided with the capability for concurrent repair of the power components and deferred repair of the cooling units.
Constant power monitoring hardware assists in detection of early loss of source power, and is designed to notify the operating system to effect an orderly shutdown. This same power monitoring hardware can detect the loss of redundant bulk power supplies, regulators, fans, and blowers and report them to the operating system for entry in the system error log and for maintenance notification.
CEC Subsystem: RAS features such as ECC on the L1 data cache and L2 data and directory caches can provide dynamic detection and correction of hard or soft array cell failures. Dynamic deallocation functions include:
- L3 cache line detect and L2 cache deallocation
- Individual L3 cache memory bypass
- Mainstore memory card deallocation
- L2 cache and CPU deallocation
- Redundant bit steering in L1 and L2 caches and L2 and L3 directory arrays
Building on the robust features of the processor complex, the pSeries 690 system has enhanced error detection and prevention capabilities. These processor RAS features are called Dynamic Processor Deallocation and Persistent Processor Deallocation. Predictive Failure Analysis® is performed on recoverable processor errors during run-time. If a processor exceeds a threshold of recoverable errors such as on L2 cache accesses, the system is designed to log the event marked and deconfigure the processor from the system while the operating system continues to run. This feature allows the processor to be repaired on a deferred basis while helping to prevent an unscheduled system outage.
Persistent Processor Deallocation is designed to ensure that on subsequent reboots, the processor remains out of the system configuration until replaced. The standard memory card has single error-correct and double-error detect ECC circuitry to correct single-bit memory failures. The double-bit detection helps maintain data integrity by detecting and reporting multiple errors beyond what the ECC circuitry can correct. The memory chips are organized such that the failure of any specific memory module should only affect a single bit within an ECC word (bit scattering), allowing for error correction and continued operation in the presence of a complete chip failure (Chipkill recovery).
This memory card also utilizes memory scrubbing and thresholding to determine when spare memory modules within each bank of memory should be used to replace ones that have exceeded their threshold value (dynamic bit steering). Memory scrubbing is the process of reading the contents of memory during idle time and checking and correcting any single-bit errors that have accumulated. These single-bit failures could be either solid (technology failures) or soft failures (intermittent errors caused by noise or other cosmic effects). If an error is detected, the system is designed to correct it by passing the data through the ECC logic that corrects the fault and then writing the corrected contents back to its memory address location.
To prevent an uncorrectable memory error from causing a system outage, the service processor is designed to initiate a deferred maintenance request on memory cards that have used their spare bits and are experiencing additional correctable errors (memory predictive failure analysis).
DASD predictive failure analysis is designed to detect an imminent disk
failure and report the findings through the AIX log. Then the disk can
be replaced either immediately or scheduled at a later time, depending on
the configuration of the DASD subsystem. With the hot-plug disk design,
normal operation can continue in the presence of a failed disk and a
concurrent repair can be performed while the application continues
uninterrupted.
Availability — UE Error Handling
Rare events in the processor, cache, and memory subsystems sometimes cause errors that exceed the error correction capability, and become Uncorrectable Errors (Ues). In other systems, these result in immediate checkstop with the offending component fully identified by the run-time diagnostics. The pSeries 690 system is designed not to cause a checkstop of the system. Instead, the corrupted data is specially marked until used by a processor, at which time a new synchronous machine check interrupt allows AIX to localize the effect to a single LPAR partition or a software process. This localizes a global system resource to affect only the partition utilizing the resource instead of checkstopping all system partitions. The affected partition can then be rebooted with significantly reduced boot time as compared to a full system reboot.
I/O Subsystem:
The interface from the processor to the I/O is through the Remote
I/O (RIO) link. This link uses a loop interconnect technology to provide
redundant paths to I/O drawers. There is a maximum of one I/O drawer on
each loop path, with a total of eight I/O drawers per system. RIO
availability features include CRC checking on the RIO bus with packet
retry on bus timeouts. In addition, if a RIO link fails, the hardware is
designed to automatically initiate a RIO bus reassignment to route the
data through the alternate path to its intended destination.
Availability — PCI Bus Error Recovery
In the pSeries 690 system, the PCI slots in the I/O expansion drawers have specialized hardware and system firmware to support PCI Bus Parity Error Recovery. The specialized hardware function is designed to interact with the software device driver to help prevent PCI slot bus errors from causing a system machine check interrupt and subsequent reboot. Instead, errors are reported locally to the specific device driver controlling activity to the PCI device. Special firmware commands allow the device driver to reset the slot and retry the command, and perform other recovery operations without affecting the total system operation.
The purpose of this design point is also to provide better fault isolation for PCI adapters because each PCI adapter is in essence on its own PCI bus. If it has an interface error, it only affects this single adapter instead of all other adapters residing on the common PCI bus. Repair time is significantly reduced because isolation is to the specific adapter and PCI hot plug allows concurrent repair of that card without affecting other adapters.
The pSeries 690 uses AIX online (concurrent) diagnostics with error log analysis and service aids. Administrators or IBM service representatives can diagnose potential system malfunction without interrupting end-user operation.
Auto-restart options, when enabled, are designed to reboot the system automatically following:
- Unrecoverable software error
- Software hang
- Hardware failure
- Environmentally-induced (ac power) failure
These standard, high-availability features, coupled with the optional
HACMP program, are designed to provide unmatched availability among UNIX
systems.
Serviceability — Correcting Faults
The pSeries 690 product is designed to be installed and maintained by a trained service representative.
Much of the repair of the pSeries 690 system can be done in either a
concurrent repair or a deferred repair scenario. Areas of the system
that support the concurrent repair scenario are PCI adapters, the entire
CEC power and cooling subsystem, the power subsystem on the I/O drawers,
and the integrated disk units. Deferred repair scenarios are supported
for the I/O drawer cooling units, RIO connections, power connections, and
predictive failure analyzed portions of the CEC including the processors
and memory subsystems when components have been varied off either
dynamically or at boot time.
Microcode Discovery Service
Microcode Discovery Service is a new service that allows you to determine if your pSeries 690 is at the latest microcode level. Using a secure Internet connection and a Web browser, Microcode Discovery Service can capture the machine data and generate a real-time comparison report showing subsystems that may need to be updated.
To accomplish this, you must install the new service tool. This tool, Inventory Scout, runs as a daemon on the server. The tool creates a file containing the current level of all microcode (adapters, devices, system, and support processor) levels in the system. Use the file to compare the system level codes against the latest available levels on the IBM Web site. You will then receive a report notifying you of any new updates available along with a link from which the microcode can be downloaded.
This service will allow you to maintain the pSeries 690 microcode at the latest levels to obtain peak performance. This service will be available through the Worldwide Support Centers.
-
Service Focal Point
To accommodate error reporting, analysis and repair in the LPAR environment, a new application was developed to run on the Hardware Management Console (HMC). This application is called the Service Focal Point (SFP) and leverages the design capabilities of the HMC to provide equivalent virtual function to the current capabilities presented by physical op panels, service processor TTY menu interfaces and system firmware interfaces and capability for configuring/reconfiguring building block hardware into partitions. The SFP is a system infrastructure that manages serviceable event information for the system building blocks. It consists of resource managers that monitor and record information about different objects in the system. It filters and correlates events from the resource managers and initiates calls home when appropriate. A user interface allows a user to view the events and perform problem determination. When a problem is corrected the user can record actions taken to resolve the hardware problem. These features of SFP support the overall problem management strategy in a complex system.
The SFP application receives service action events from the Service Processor for critical system down situations, and from the Service Agent application programs running on the individual logical partitions for system recoverable or predictive events and Operating System or device driver detected events.
-
Service Action Event Log
The SFP collects the serviceable events from different building blocks together in a Service Action Event (SAE) log. The log entries are generated by analysis routines that run on an error in a building block. The resource manager for the building block forwards information about the event to the service focal point and the information is placed in the SAE log. The particular content of the error data depends upon the type of the error and on the system configuration itself.
The SAE log on the SFP also contains pointers to extended information that may have been recorded at the time of a serviceable event by the building block. Extended error collection includes not only the collection of first failure data capture, but also vital product data, partition information, operating system error logs, service processor error logs, error register data, and so forth.
When the SFP receives a new log entry, filtering determines if this was a unique event. Sometimes an event notification can come from more than one resource manager for the same event or a resource manager may forward a notification for an event which previously occurred but has not been corrected yet.
-
Service Agent Component
When a service action event is logged in the SFP, the system needs to communicate the failure back to IBM. During this call-home function, particular error data and system configuration information needs to be sent to IBM to drive the service delivery infrastructure. The SFP utilizes the Service Agent focal point application residing on the HMC along with the HMC modem to initiate the call home and transfer the pertinent error information to IBM Service. When a call-home is required, Service Agent manages the connection to IBM, which is used to open a problem record. The problem record is used by the service delivery team to determine whether or not to dispatch a customer engineer with the appropriate service parts to the system to perform a repair.
When service personnel perform a repair on the system, the SFP is used to identify the source of the problem and record information relating to the repair. When the repair is complete, the service representative can update the SAE log with FRU replacement information and any comments that the service representative has. The information stored by the SFP represents the system's service history and can be used to ensure proper maintenance over the life of the system.
-
Guiding Light Maintenance
To assist the service representative in locating the correct system unit and drawer that contain the fault requiring repair, the SFP can enable the capability to flash LEDs on the respective system unit and drawer that contain the fault.
Service Support
Hardware service requests go to IBM's remote support center for initial
problem diagnosis. This approach gives you direct access to skilled
specialists. These specialists can either solve the problem over the
phone or help get it resolved as quickly as possible. They are backed up
by a highly-trained product engineering team to assist in problem
identification and resolution.
Euro Currency
These products are not impacted by euro currency.
Product Positioning
The pSeries 690 will benefit customers in all segments. Traditional enterprise customers need more powerful systems as they integrate and extend their mission-critical operations, such as Enterprise Resource Planning (ERP), Customer Relationship Management (CRM), and Supply Chain Management (SCM) to include e-business and e-commerce, the subset of e-business concerned with buying and selling using the Internet. It provides the availability, reliability, scalability, and performance that address key unmet customer needs:
- Compete effectively by increasing business efficiency
- Streamline decision management by integrating business and management processes
- Improve customer service by extending business and management processes to include customers and business partners
This UNIX high-end mainframe-like superserver is ideal for large enterprises with heavy duty commercial workloads that require highly reliable continuous operation for scientific and technical workloads that require significant number crunching and near linear scalability. It has the flexibility to perform as a high-end SMP with LPAR and HACMP high-availability software.
The benefits of the pSeries 690 vary by usage. For server consolidation, reducing the complexity of managing systems, reducing the physical footprint, increasing flexibility, improving availability, increasing security, and decreasing overhead costs can be realized by using system capabilities.
For the e-Infrastructure environments, quality of service to customers is enhanced, level of system security is increased, the time to deploy is decreased, the infrastructure is more easily managed, performance is maximized, and new technologies are integrated and utilized.
The pSeries 690 also provides an excellent approach to a wide range of high-performance computing applications because of its high-performance characteristics, quality operating environment, wide range of configurations, and ability to be clustered as workload demands increase. pSeries 690 systems configured with 4-way, HPC option processor Multi-Chip Modules (MCM features #5240 and #5250) may be especially attractive to customers in the high performance computing environment. This configuration provides twice the amount of Level 2 and Level 3 cache, as well as additional memory bandwidth per processor when compared to pSeries 690 systems configured with 8-way processor MCMs. The additional cache and memory bandwidth provided by this configuration may provide significantly higher performance per processor for certain engineering and technical environment applications.
The fast engine and increased performance, the ability to tackle larger problems, leadership technology and skill management provide the benefits in a high-performance computing environment.
In addition, for core applications, the pSeries 690 enables additional growth for database engines, high application availability, two-tier approaches using LPAR, and increased BI enablement due to performance and capacity improvements.
Statement of General Direction
IBM is planning an exciting new IBM Capacity Advantage offering
for the pSeries 690 that allows for rapid dynamic activation of installed
inactive processors and memory to existing partitions without disrupting
applications. This feature allows for increased flexibility to respond
to changing workload requirements. In the unlikely event of a failure,
inactive resources can also be used to replace a failing component, again
without disruption.
This offering is an extension of the current Capacity Advantage program
available on IBM . It is planned for introduction on the pSeries 690
in second half 2002. pSeries 690 systems purchased before then can
participate fully along with new systems.
IBM is also planning to enable you to dynamically redefine the resources that makeup a logical partition. This means that partitions can be changed while applications continue to run. No reboot of affected partitions will be required after this feature is installed. Processors, memory, and I/O can all be added or removed from active partitions without disruption to applications. This facility is planned for introduction on the pSeries 690 in second half 2002, and will be supported by an update to AIX and the system software.
All statements regarding IBM's future direction and intent are subject to change or withdrawal without notice, and represent goals and objectives only.
Reference Information
For additional information on SP and PSSP cluster configuration support on pSeries 690 servers, refer to Hardware Announcement 101-283 , dated October 4, 2001.
Trademarks
Publications
The following publications are shipped with the IBM
pSeries 690. Additional copies are available. To order, contact your
IBM representative.
Order
Title Number
pSeries 690 Installation Guide SA38-0587
System Unit Safety Information SA23-2652
PCI Adapter Placement Reference SA38-0538
Refer to the Related Publications section of the pSeries 690 Installation Guide for information regarding other available pSeries 690 publications or on the Web, visit:
The following publications are available for a fee:
Order
Title Number
pSeries 690 Installation Guide SA38-0587
pSeries 690 User's Guide SA38-0588
pSeries 690 Service Guide SA38-0589
HMC Operations Guide SA38-0590
HMC Maintenance Guide SA38-0603
Diagnostic Information for MultiBus SA38-0509
Systems
Adapters, Devices and Cable Information SA38-0516
PCI Adapter Placement Reference SA38-0538
Site Hardware Planning Information SA38-0508
The Publication Notification System (PNS) is available by order number/product number. Customers currently subscribing to PNS will receive notifications by e-mail automatically. Customers who want to subscribe can visit the PNS Web site location at:
The publications listed on the notification can be ordered by calling the Publications Support Group in Raleigh at 800-879-2755, option 1.
The IBM Publications Center Portal:
The Publications Center is a worldwide central repository for IBM product publications and marketing material with a catalog of 70,000 items. Extensive search facilities are provided, as well as payment options via credit card. Many publications are available online in various file formats and can be downloaded free of charge.
Note: PNS subscribers most often order their publications via the Publication Center.
Technical Information
pSeries 690 (7040-681) (single rack)
-
Width:
- 79 cm (30.9 in) (with acoustic rear door)
- 79 cm (30.9 in) (with slim-line rear door)
-
Depth:
- 144 cm (56.8 in) (with acoustic rear door)
- 134 cm (52.8 in) (with slim-line rear door)
-
Height:
- 202 cm (79.5 in) (with acoustic rear door)
- 202 cm (79.5 in) (with slim-line rear door)
-
Weight:
- 1,209 kgs (2,666 lbs maximum) (with acoustic rear door)
- 1,196 kgs (2,636 lbs maximum) (with slim-line rear door)
pSeries 690 (7040-681) (with Expansion Rack feature)
-
Width:
- 154 cm (60.6 in) (with acoustic rear door)
- 154 cm (60.6 in) (with slim-line rear door)
-
Depth:
- 144 cm (56.8 in) (with acoustic rear door)
- 134 cm (52.8 in) (with slim-line rear door)
-
Height:
- 202 cm (79.5 in) (with acoustic rear door)
- 202 cm (79.5 in) (with slim-line rear door)
-
Weight:
- 1,850 kgs (4,074 lbs maximum) (with acoustic rear door)
- 1,823 kgs (4,019 lbs maximum) (with slim-line rear door)
- Temperature: 16° to 32°C (61° to 90°F)
- Relative humidity: 8% to 80%
- Maximum wet bulb: 23°C (73°F)
-
Sound power:
- 7.4 bels (operating/idling with acoustic covers)
- 8.0 bels (operating/idling with slim-line covers)
-
Sound pressure:
- 58 dBa (operating/idling with acoustic covers)
- 66 dBa (operating/idling with slim-line covers)
-
Maximum altitude:
- 1.1 GHz Processors: 3,040 m (10,000 ft)
- 1.3 GHz Processors: 2,134 m (7,000 ft)
-
Operating voltage at 50 to 60 Hz: (three-phase)
- 200 to 240 V ac
- 380 to 415 V ac
- 480 V ac
- Electrical output: 15,400 watts (maximum)
- Power source loading: 15.0 kVA (maximum configuration)
- Thermal output: 14,910 joules/sec (51,200 Btu/hr) (maximum configuration, 32-way CEC)
EMC Conformance Classification
This equipment is subject to FCC rules and it shall comply with the appropriate FCC rules before final delivery to the buyer or centers of distribution.
- U.S. — FCC CFR47 Part 15 Class A
- Europe — CISPR 22 Class A; "CE" Mark of Conformity
- Japan — VCCI-1
- Korea — Korean Requirement Class A
Homologation: Telecom Environmental Testing (Safety and EMC): Homologation approval for specific countries has been initiated with the IBM Homologation and Type Approval (HT&A) organization in LaGaude.
This IBM pSeries model and applicable features meet the environmental testing requirements of the country TELECOM and have been designed and tested in compliance with the Full Quality Assurance Approval (FQAA) process as delivered by the British Approval Board for Telecom (BABT), the U.K. Telecom Regulatory authority.
Product Safety/Country Testing/Certification
- U.S.: UL
- Canada: CNL (CSA or cUL)
- Germany/Europe: GS Mark (Safety, TUV, EN60 950)
General Requirements
The pSeries 690 is in compliance with IBM Corporate Bulletin C-B
0-2594-000 Statement of Conformity of IBM Product to External Standard
(Suppliers Declaration).
Hardware Requirements
pSeries 690 — Minimum System Configuration: The minimum configuration for an pSeries 690 includes the following items:
-
One 7040-681 Central Electronics Complex (17U rack space required), and
one each of the following (except as noted):
- CD-ROM (#2624)
- Diskette Drive Cable (#3155)
- Operator Panel Attachment Cable (#3255)
- 128 MB Level 3 Cache (4 X 32 MB), 400 MHz (#4138)
- 8 GB Memory Card, Inward Facing (#4181)
- 8-Way, POWER4 Processor (First) (#5242)
- Processor Clock Card, Programmable (#5251)
- Power Cable Group, Bulk Power to CEC and Fans (#6161)
- Interface Cable, Power to Service Processor (#6162)
- Two Power Converter Assembly, Central Electronics Complex (#6170)
- Power Cable Group, CEC to Power Controller, First Processor Module (#6181)
- Capacitor Book, First and Second Processor Modules (#6198)
- Service Processor with Remote I/O Loop Attachment, Two Loops (#6404)
- Backplane, Central Electronics Complex (#6565)
- Media Drawer, Operator Panel, Diskette (#8692) — (1U Rack Space Required)
- Drawer Placement Indicator, 18U Position, Primary Rack (#4618)
- Language Specify (#9xxx)
-
One 7040-61D I/O Drawer (4U rack space required)
- One SCSI Cable — I/O Drawer to Media Drawer (#2122)
- One Remote I/O Cable, 05.M (#3145) (between drawer halves)
- Two Remote I/O Cables, 2M (#3149)
- Two 18.2 GB, 10,000 RPM Ultra3 SCSI Disk Drive Assembly (#3157)
- One I/O Drawer Attachment Cable Group, Drawer Position #4605 (#6121)
- Two Power Converter Assembly, I/O Drawer (#6172)
- One Power Cable, I/O Drawer to Media Drawer (#6179)
- One PCI Single Ended SCSI Adapter (Media Drawer Attach) (#6206)
- Two I/O Drawer PCI Planar, 10 Slot, 2 Integrated Ultra3 SCSI Ports (#6563)
- Four Ultra3 SCSI 4-Pack Hot-Swap Back Plane (#6564)
- One Drawer Placement Indicator, 9U Position, Primary Rack (#4609)
- One Language Specify (#9xxx)
-
One 7040-61R System Rack (24-inch rack with 42U available space)
- One Front Door, Black with Copper Accent, Primary Rack (#6070)
- One Rear Door, Slim Line, Primary or Secondary Rack (#6074)
- Two Bulk Power Regulators (#6186)
- Two Power Controller, Four Cooling Fans, Three dc Power Converter Connections (#6187)
- Two Power Distribution Assembly, 10 dc Power Converter Connections (#6188)
- One Bulk Power Assembly, Redundant (#8690)
- One Rack Content Specify, 7040/681 — 17U (#0190)
- One Rack Content Specify, 7040/61D — 4U (#0191)
- One Rack Content Specify, 7040/681 Feature #8692 — 1U (#0192)
- One Rack Content Specify 7040/61R Feature #8490 — 8U (#0193)
- Two Line Cords, 60 A/240 V ac, 6-AWG, 14 feet, IEC309 Plug (#8678)
- One Language Specify (#9xxx)
Software Requirements
- AIX® 5L for POWER Version 5.1 with the 5100-01 Recommended Maintenance package (APAR IY21957), or later
pSeries 690 Central Electronics Complex (CEC) (7040-681)
- The pSeries 690 CEC (7040-681) is a 17U tall, 24-inch, rack-mounted device. It houses the system processors, memory, system support processor, I/O drawer connection capability, and associated components. It is installed with the bottom of the drawer at placement indicator (#4418), the 18U position of the 7040-61R rack.
-
The pSeries 690 is powered by one to four multichip processor modules.
Each Multichip Module (MCM) contains either four or eight processors.
The pSeries 690 is available in the following configurations:
- 8- and 16-way 1.3 GHz, POWER4, HPC option processors (4 processors per MCM)
- 8-, 16-, and 24-way 1.1 GHz, POWER4 processors (8 processors per MCM)
- 16-, 24-, and 32-way 1.3 GHz, POWER4, Turbo option processors (8 processors per MCM)
- All processors in a pSeries 690 must operate at the same speed.
- Each processor contains 32 KB of data cache and 64 KB of instruction cache.
- Each processor MCM contains by 5.6 MB of Level 2 cache which is shared between the processors. This provides 1.44 MB of L2 cache per processor for four processor MCMs and 0.72 MB of L2 cache per processor for eight processor MCMs.
-
Each installed processor MCM is supported with 128 MB of Level 3
cache. This provides 32 MB of L3 cache per processor for four processor
MCMs and 16 MB of L3 cache per processor for eight processor MCMs. The
L3 cache operates at 400 MHz for 1.1 GHz processors (#4138) and 433 MHz
for 1.3 GHz processors (#4139).
Note: pSeries 690 systems configured with 4-way, HPC option processor MCMs (features #5240 and #5250) may be especially attractive to customers in the high performance computing environment. This configuration provides twice the amount of Level 2 and Level 3 cache, as well as additional memory bandwidth, per processor when compared to systems configured with 8-way processor MCMs. This additional cache and memory bandwidth available for each processor in this configuration may provide significantly higher performance per processor for certain engineering and technical environment applications.
- A programmable processor clock card (#5251) is required to provide system and processor clocking.
- pSeries 690 systems configured with two or three processor MCMs must have the empty processor positions populated with Processor Bus Pass Through Modules (#5257). No Processor Bus Pass Through Modules are required on systems with one processor MCM.
- pSeries 690 processor MCMs require capacitor books to operate. One capacitor book (#6198) is required for each two populated processor MCM positions.
- The pSeries 690 utilizes redundant power. It implements redundant bulk power assemblies, bulk power regulators, power controllers, power distribution assemblies, DC power converters and associated cabling.
- Power for the pSeries 690 CEC is supplied from dc bulk power assemblies in the 7040-61R rack. The bulk power is converted to the power levels required for the CEC via dc power convertors. Two dc power converters (#6170) and attachment cables (#6181 and #6161) are required for the CEC and the first processor MCM, memory, and miscellaneous CEC components. Each additional MCM (second, third, or fourth) requires an additional dc power converter (#6189) and the appropriate power cable group (#6182, 6183, or 6184).
- Up to eight 7040-61D I/O drawers can be attached to the pSeries 690. The Support Processor and Remote I/O Loop Attachment card (#6404) provides system control function and four remote I/O ports for attachment of up to two 7040-61D I/O drawers via separate RIO loops. Each Remote I/O Loop Adapter (#6410) provides an additional eight RIO ports supporting allowing the attachment of up to four additional 7040-61D I/O drawers.
-
Two RIO Loops (four RIO ports) are activated for each populated processor
MCM.
- One MCM allows attachment of up to two I/O drawers
- Two MCMs allow attachment of up to four I/O drawers
- Three MCMs allow attachment of up to six I/O drawers
- Four MCMs allow attachment of up to eight I/O drawers
- One Media Drawer (#8692) is required for each pSeries 690 server. The media drawer is a 1U rack drawer located directly below the CEC at the 17U location of the 7040-61R rack. It incorporates the system operator panel, a diskette drive, and up to four media devices.
- The media drawer is split into two separate sections. The front section houses the system operator panel, a diskette drive, and two media bays. Utilization of the rear section is optional to provide space for two additional media devices if desired.
- The front and rear sections of the media drawer must be connected to separate SCSI PCI adapters located in one of the systems I/O drawers. When connecting to a SCSI adapter with a 68 pin P-style connector the I/O drawer to media drawer SCSI cable #2122 is used. When connecting to a SCSI adapter with a Mini-68 pin VHDCI connector converter cable #2118 is used in conjunction with #2122.
- Power for the media drawer is provided by the I/O drawer.
- One media device capable of reading CD-ROM media is required for each pSeries 690 system.
- DVD-RAM (#2623) and 4 mm tape (#6158) devices are allowed only in the front bays of the media drawer.
- U.S. English (#8800 and #8840) USB keyboards can be ordered with the 7040-61D I/O Drawer for use with native-attached displays on the pSeries 690.
-
Various native language groups are supported by AIX on pSeries 690.
Keyboard function with the following languages has been verified. USB
keyboards for these languages can be obtained anywhere IBM products are
sold such at the IBM Web site. For more information on where to buy
these IBM products, contact your local IBM representative.
- U.K. English, #166, Black USB Keyboard
- French, #189, Black USB Keyboard
- German/Austrian, #129, Black USB Keyboard
- Italian, #142, Black USB Keyboard
- Japanese, #194, Black USB Keyboard
- Russian, #443, Black USB Keyboard
- Spanish, #172, Black USB Keyboard
- L.A. Spanish, #171, Black USB Keyboard
- Chinese/U.S., #467, Black USB Keyboard
-
The pSeries 690 has eight memory slots. Four memory slots utilize inward
facing memory cards and four utilize outward facing memory cards. The
inward facing memory slots are utilized by the first and second processor
module positions while the outward facing memory slots support the third
and fourth 8-way processor module positions.
-
A pSeries 690 system with only one processor MCM position populated has
access to only two inward facing memory slots.
-
Minimum system memory is 8 GB. Maximum system memory is 256 GB.
-
Memory is available in 4, 8, 16, and 32 GB inward facing and outward
facing increments.
-
The following memory features must be ordered and installed in identical
pairs:
- 4 GB Memory Card, Inward Facing (#4196)
- 4 GB Memory Card, Outward Facing (#4197)
- 16 GB Memory Card, Inward Facing (#4183)
- 16 GB Memory Card, Outward Facing (#4184)
-
All memory cards in pSeries 690 systems equipped with either 4 GB or 16
GB memory cards must be of the same capacity.
-
Memory placement and balancing across processors is desirable for certain
workloads. This is especially important in the technical computing
environment where maximum memory bandwidth is desired. In such cases,
consideration should be given to balancing memory across all populated
processor module positions to obtain the best performance. Unbalanced
configurations will function properly, but may not provide optimum
performance for the planned applications.
Example: In many technical environment applications, a system with four processor module positions populated and 64 GB of memory should perform better with four 16 GB features installed than with two 32 GB features installed. In many applications, utilizing eight 8 GB memory cards may even further enhance performance. Plans for future memory upgrade requirements should always be balanced against projected performance needs to ensure desired future memory expansion is not unnecessarily limited.
For additional information regarding the effect of memory card placement and balancing on system performance review the Configuring pSeries 690 for Performance white paper.
- IBM recommends systems configured with 4 GB of memory, or greater, to have access to a 4- or 8-mm tape drive for submission of system dump information, if required. This function may be accomplished via locally attached or network attached devices as appropriate.
- The pSeries 690 can be divided into as many as 16 logical partitions. System resources can be dedicated to each LPAR.
- LPAR allocation, monitoring, and control is provided by the Hardware Management Console (#7315).
- Each LPAR functions under its own instance of the operating system.
- A minimum of one processor is required per LPAR.
- A minimum of 4 GB of system memory is recommended per LPAR.
- I/O adapters in PCI slots may be allocated to an LPAR on an individual slot basis.
- Integrated Ultra3 SCSI controllers located in the 7040-61D I/O drawers may be individually allocated to an LPAR. These integrated SCSI adapters each support one 4-pack disk backplane.
- While it is not mandatory, consider allocating one half of a 7040-61D I/O drawer for each LPAR. This would provide one 10-slot PCI planar, two integrated SCSI controllers, and two 4-Pack SCSI disk backplanes to the LPAR. This will help to ensure balanced I/O bandwidth for the LPAR configurations.
IBM Hardware Management Console for pSeries
-
The IBM Hardware Management Console for pSeries (HMC) (#7315) provides a
set of functions necessary to manage the pSeries 690 and LPAR
configurations. These functions include:
- Create and store LPAR profiles that define the processor, memory, and I/O resources allocated to an individual partition
- Starting, stopping, and resetting a system partition
- Boot a partition or system by selecting a profile
- Display system and partition status
- Displaying the contents of a virtual operator panel for each partition or controlled system
- The HMC also provides a service focal point for the systems it controls. It is connected to the service processor of the system via a dedicated serial link. The HMC provides tools for problem determination and service support, such as call-home and error log notification through an analog phone line.
- The HMC is a dedicated function device. It is utilized only for the control and service functions of the pSeries servers it serves. It is not available for use as a general purpose computing resource.
- Each HMC must be configured with a graphics display and a keyboard. A mouse is included with each HMC.
- One HMC can control up to four pSeries servers.
- If redundant HMC function is desired, the pSeries 690 can be attached to two separate hardware management consoles.
-
An Ethernet connection between the HMC and each active partition on the
pSeries server is recommended. This connection is utilized to provide:
- Additional systems management such as WebSM management of AIX in the individual partitions
- Collection and passing of hardware service events to the HMC for automatic notification of error conditions to IBM
- Total system inventory collection
- The two integrated serial ports on the HMC may be utilized for connection to the pSeries server. If additional serial ports are required, 8-port asynchronous adapters (#2943) should be utilized.
- Various native language keyboards are available for the hardware system console. However, the HMC displays menus and control information in English regardless of the keyboard utilized.
- Each pSeries 690 CEC must be connected to a #7315 Hardware Management Console (HMC) for system control, LPAR (logical partitioning) and service functions. The HMC is capable of supporting multiple pSeries 690 systems if desired.
- The pSeries 690 provides two special async ports especially dedicated for HMC attachment. The connectors on these ports are differentiated from standard async ports by providing a 9-pin female connector vs the standard 9-pin male connector. The connection between the HMC and the pSeries 690 CEC is through cables available in 6-meter (#8120) and 15-meter (#8121) lengths.
- The pSeries 690 provides two async ports for attachment of ASCII terminals and devices. These two async ports are under the control of the system's service processor and are not available for functions such as HACMP heartbeat cabling or UPS control which require fully dedicated ports. The service processor menus may be accessed through an ASCII terminal attached to one of the async ports or via an internal firmware controlled "wrap" connection to the HMC. When not in use for accessing service processor menus, the async ports are available for general purpose uses that do not require dedicated ports.
- The pSeries 690 utilizes the 7040-61D I/O drawer for directly attached PCI adapters and SCSI disk capabilities. The pSeries 690 requires a minimum of one 7040-61D drawer; a maximum of 8 I/O drawers can be attached.
- Each 7040-61D I/O drawer is divided into two separate halves. Each half contains ten blind-swap PCI slots and two Ultra3 SCSI 4-pack backplanes for a total of twenty PCI slots and sixteen hot-swap disk bays per drawer.
- One single-wide PCI blind-swap cassette (equivalent to those in feature #4599) is provided in each PCI slot of the 7040-61D. Cassettes not containing a PCI adapter will be shipped with a "dummy" card installed to ensure proper environmental characteristics for the drawer. If additional single-wide blind-swap cassettes are needed, order feature #4599.
- All ten PCI slots on each I/O drawer planar can support either 64-bit or 32-bit PCI adapters. Each I/O drawer planner provides seven 3.3V, 66 MHz, 64-bit slots and three 5V, 32 MHz, 64-bit slots.
- Each I/O drawer planar incorporates two integrated Ultra3 SCSI adapters for direct attachment of the two 4-pack hot swap backplanes in that half of the drawer. These adapters do not support external SCSI device attachments.
- I/O drawers are connected to the pSeries 690 CEC via RIO loops. RIO connections are always made in loops to help protect against a single point-of-failure resulting from an open, missing, or disconnected cable. pSeries 690 systems with non-looped configurations could experience degraded performance and serviceability. Two 2-meter RIO cables (#3149) are used between each half of the I/O drawer and the CEC. The separate drawer halves are connected via a 0.5 meter RIO cable (#3145) to complete the loop.
- Each half of the 7040-61D I/O drawer is powered separately.
-
I/O drawers occupy specific locations in the 7040-61R rack or expansion
feature rack. Drawer placement is specified by the mounting location of
the bottom of the drawer. Drawer placement is as follows:
- Drawer #1 — Placement indicator #4609 — 9U position of 7040-61R rack
- Drawer #2 — Placement indicator #4605 — 5U position of 7040-61R rack
- Drawer #3 — Placement indicator #4601 — 1U position of 7040-61R rack
-
Drawer #4:
- Placement indicator #4613 — 13U position of 7040-61R rack if Integrated Battery Backup feature (#6200) is not installed.
- Placement indicator #4401 — 1U position of expansion rack if Integrated Battery Backup feature (#6200) is installed at location #4613.
-
Drawer #5:
- Placement indicator #4401 — 1U position of expansion rack if Integrated Battery Backup feature (#6200) is not installed.
- Placement indicator #4405 — 5U position of expansion rack if Integrated Battery Backup feature (#6200) is installed at location #4613
-
Drawer #6:
- Placement indicator #4405 — 5U position of expansion rack if Integrated Battery Backup feature (#6200) is not installed.
- Placement indicator #4409 — 9U position of expansion rack if Integrated Battery Backup feature (#6200) is installed at location #4613.
-
Drawer #7:
- Placement indicator #4409 — 9U position of the expansion rack if Integrated Battery Backup feature #6200 is not installed.
- Placement indicator #4413 — 13U position of the expansion rack if Integrated Battery Backup feature #6200 is installed at location #4613.
-
Drawer #8:
- Placement indicator #4413 — 13U position of the expansion rack if Integrated Battery Backup feature #6200 is not installed.
- Placement indicator #4419 — 19U position of the expansion rack if Integrated Battery Backup feature #6200 is installed at location #4613.
- Minimum of two internal SCSI HDDs are required per pSeries 690 server. These disks should be utilized as mirrored boot devices and should be mounted in the first 7040-61D I/O drawer. This configuration provides service personnel the maximum amount of diagnostic information if the system encounters errors in the boot sequence.
- Boot support is also available from local SCSI, SSA, and Fibre Channel adapters, or from network via ENET or token-ring adapters.
- Consideration should also be given to the placement of AIX rootvg volume group in the first I/O drawer. This allows AIX to boot any time other I/O drawers are found offline during boot.
- If the boot source other than internal disk is configured, the supporting adapter should also be in the first I/O drawer.
- The pSeries 690 incorporates an Early Power Off Warning (EPOW) capability that assists in performing an orderly system shutdown in the event of a sudden power loss. IBM recommends use of the integrated battery backup features or an uninterruptible power system (UPS) to help ensure against loss of data due to power failures.
- The 7040-61R rack is a 24-inch rack with an integrated power subsystem to support the pSeries 690 system. It provides 42U of rack space.
- An expansion rack feature (#8691) is available if additional 24-inch rack space is required. To install the expansion rack feature the side cover of the 7040-61R rack is removed, the expansion rack is connected to the 61R, and the side cover is placed on the exposed side of the feature rack. Power for components in the expansion rack is provided from the bulk power assemblies in the Model 61R rack.
- Over time, the 7040-61R rack will be utilized in multi-rack clusters. Rack indicators (such as #4651) will be used to identify the racks within the cluster and specify the desired rack in which to place various rack mounted components. When assigning rack indicators, the 7040-61R and its expansion rack feature #8691 are considered a single rack.
-
All 7040-61R racks and expansion feature racks must have door assemblies
installed. The following door assemblies are available:
- A sculptured black front door with copper accent (#6070) is required for the primary Model 61R rack.
- A sculptured black front door (#6071) is required for the expansion rack feature, if installed.
- An acoustic rear door (#6075) is available for the primary rack or the expansion feature rack. This door should be utilized for installations requiring a quieter environment.
- A slim-line rack door (#6074) is available for the primary or expansion rack feature. This door should be utilized for installations where system footprint is the primary consideration.
- If additional external communication and storage devices are required, 7014-T00 or T42 racks should be ordered. There is no limit on the quantity of 7014 racks allowed.
- The height of the 7040-61R rack (42U) or expansion feature rack may require special handling when shipping by air or when moving under a low doorway.
- The 7040-61R rack always incorporates two bulk power assemblies for redundancy (#8690). These provide 350 V dc power for devices located in the 61R rack and associated expansion rack feature. These bulk power assemblies are mounted in front and rear positions and occupy the top 8U of the rack. To provide optimum system availability these bulk power assemblies should be powered from separate power sources via separate line cords.
- Bulk power regulators (#6186) interface to the bulk power assemblies to ensure proper power is supplied to the systems components. The number of bulk power regulators required is configuration dependent, based on the number of processor MCMs and I/O drawers installed. The following table provides the number of bulk power regulators required for each combination of processor MCMs and I/O drawers.
Number of Processor
MCMs Installed
1 2 3 4
1 2 2 4 4
2 2 2 4 4
Number of 7040-61D 3 2 4 4 4
I/O Drawers Installed 4 2 4 4 4
5 4 4 4 4
6 4 4 4 6
7 4 4 4 6
8 4 4 6 6
- Two Power Controller features (#6187) are required for all pSeries 690 systems, one per bulk power assembly. Each Power Controller provides power connections to support the systems four cooling fans. It also provides three additional connectors that can be utilized to attach dc power converters contained in the CEC and each I/O drawer. Ten additional connector locations are provided by the Power Distribution Assembly (#6188). Two Power Distribution Assemblies (one per Bulk Power Assembly) are required on each system. Two additional Power Distribution Assemblies may be added as required to provide more connections for large system configurations.
- Optional Integrated Battery Backup is available, if desired. The primary and secondary battery backup features (#6200 and #6201) interface with the power controllers of the front and rear bulk power assemblies. The battery backup features protect against power line disturbances and provide sufficient power to allow an orderly system shutdown in the event that the power sources fail. The battery backup features require 4U of rack space and are always installed at the 13U location of the primary 7040-61R rack.
- System maximum limits for adapters and devices may not provide optimal system performance. These limits are given for connectivity and functionality assurance.
- Configuration limitations help ensure appropriate PCI bus loading, adapter addressing, and system and adapter functional characteristics when ordering I/O drawers. These I/O drawer limitations are in addition to individual adapter limitations shown in the feature descriptions section.
-
Combined quantities of the following adapters are limited to 40 adapters
per pSeries 690 server, 10 per 7040-61D I/O drawer, and 5 per PCI planar
(#6563):
- TURBOWAYS® 622 Mbps PCI MMF ATM Adapter (#2946)
- Gigabit Ethernet — SX PCI Adapter (#2969)
- 10/100/1000 Base-T Ethernet PCI Adapter (#2975)
- PCI Dual Channel Ultra3 SCSI Adapter (#6203)
- Gigabit Fibre Channel Adapter for 64-bit PCI Bus (#6228)
-
Advanced SerialRAID Plus Adapter (#6230)
Note: This rule does not apply when only Advanced SerialRAID Plus Adapters (#6230) are installed on a pSeries 690 configuration with none of the other listed adapters installed. In a configuration which does not incorporate any of these other adapters, 48 adapters (#6230) are allowed per pSeries 690, 12 per 7040-61D I/O drawer, and 6 per planar (#6563).
- Most PCI adapters for the pSeries 690 system can be hot-plugged. Any PCI adapter supporting a boot device or system console should not be hot-plugged.
Software Publications and Media Delivery
-
Single Pallet Delivery of software publications and media will not be
offered on the pSeries 690. As a result, to order software preinstalled
on a pSeries 690, the following things must be done.
- The hardware order must contain feature #5005.
- The 5692-A5L order must contain both feature #5005 and #3506.
- No other software order associated with the system can contain either feature #5005 or #3483.
- For additional information on ordering AIX and related software products, review AIX announce documents.
Planning Information
Cable Orders:
No additional cables are required.
Security, Auditability, and Control
This product uses the security and auditability features of AIX 5L for POWER Version 5.1 operating system.
The customer is responsible for evaluation, selection, and implementation of security features, administrative procedures, and appropriate controls in application systems and communications facilities.
Terms and Conditions
This product is available for purchase under the terms of the IBM Customer Agreement (ICA).
IBM hardware products are manufactured from new parts, or new and used parts. In some cases, the hardware product may have been previously installed.
Regardless, IBM warranty terms apply.
Volume Orders: Contact your IBM representative.
IBM Credit Corporation Financing: Yes
Warranty Service: IBM On-Site Repair (IOR)
IBM Hourly Service Rate Classification: Two
When a type of service involves the exchange of a machine part, the replacement may not be new, but will be in good working order.
Field-Installable Features: Yes
Licensed Machine Code: IBM licensed machine code is licensed for use by a customer on a specific machine, designated by serial number, under the terms and conditions of the IBM Agreement for Licensed Machine Code.
Educational Allowance: A reduced charge is available to qualified education customers. The educational allowance may not be added to any other discount or allowance.
Prices
The prices provided in this announcement are suggested retail prices for the U.S. only and are provided for your information only. Dealer prices may vary, and prices may also vary by country. Prices are subject to change without notice. For additional information and current prices, contact your local IBM representative.
Model Purchase
Description Number Price MMMC(1)
Machine Type: 7040
I/O Drawer 61D $ 3,500 $194
System Rack 61R 5,500 188
pSeries 690 CEC 681 17,789 985
Initial/
Model Feature Purchase MES/
Description Number Number Price Both
Rack Content Specify
7040/681 -- 17U 61R 0190 N/C Initial
7040/61D -- 4U 61R 0191 N/C Initial
7040/681 Feature #8692 -- 1U 61R 0192 N/C Initial
7040/61R Feature #8690 -- 8U 61R 0193 N/C Initial
CCS Customer Service 61D 0986 N/C Initial
Specify (U.S.) 61R Initial
681 Initial
PCI SCSI Adapter 16-bit 61D 2114 $ 270 Both
Differential External
Y Cable
PCI SCSI Adapter To 61D 2116 120 Both
2-Port, 16-Bit
Differential Ext
Device Cable
Converter Cable, VHDCI to 61D 2118 60 Both
P, Mini-68 pin to 68 pin,
0.3M
SCSI Cable -- I/O Drawer 61D 2122 275 Both
to Media Drawer
0.6M 16-bit SCSI-2 System 61D 2424 108 Both
to System Cable
2.5M 16-bit SCSI-2 System 61D 2425 130 Both
to System Cable
LC-SC Fibre Channel 61D 2456 100 Both
Converter Cable
4.7 GB SCSI-2 DVD-RAM 681 2623 1,200 Both
Drive (Black Bezel)
32x (Maximum) SCSI-2 681 2624 375 Both
CD-ROM Drive
IBM Short-wave Serial 61D 2732 12,000 Both
HIPPI PCI Adapter for
RS/6000(R)
IBM Long-wave Serial HIPPI 61D 2733 12,000 Both
PCI Adapter for RS/6000
Keyboard/Mouse Attachment 61D 2737 300 Both
Card -- PCI
SysKonnect SK-NET
FDDI-LP SAS PCI 61D 2741 1,590 Both
FDDI-LP DAS PCI 61D 2742 2,500 Both
S/390(R) ESCON(R) 61D 2751 9,500 Both
Channel PCI Adapter
POWER GXT135P Graphics 61D 2848 495 Both
Accelerator
Asynchronous Cable 61D 2936 73 Both
EIA-232/V.24
8-Port Asynchronous 61D 2943 1,395 Both
Adapter EIA-232/RS-422, 681 Both
PCI bus
128-Port Asynchronous 61D 2944 1,295 Both
Controller, PCI bus
Turboways 622 Mbps PCI 61D 2946 3,995 Both
MMF ATM Adapter
Cable
V.24 / EIA-232 61D 2951 175 Both
V.35 61D 2952 320 Both
V.36 / EIA-499 61D 2953 255 Both
X.21 61D 2954 175 Both
Initial/
Model Feature Purchase MES/
Description Number Number Price Both
2-Port Multiprotocol 61D 2962 $ 2,000 Both
PCI Adapter
Gigabit Ethernet -- SX 61D 2969 2,125 Both
PCI Adapter
10/100/1000 Base-T 61D 2975 1,600 Both
Ethernet PCI Adapter
Serial to Serial Port 61D 3125 80 Both
Cable for Rack/Rack
Remote I/O Cable
0.5M 61D 3145 310 Both
2M 61D 3149 350 Both
Diskette Drive Cable 681 3155 150 Both
18.2 GB 10,000 rpm Ultra3 61D 3157 1,500 Both
SCSI Disk Drive Assembly
36.4 GB 10,000 rpm Ultra3 61D 3158 2,600 Both
SCSI Disk Drive Assembly
Operator Panel Attachment 681 3255 100 Both
Cable
Color Monitor, Stealth Black
P76, Captured Cable 61D 3627 995 Both
P260 and Cable 61D 3628 1,995 Both
Color Monitor, Pearl White
P76, Captured Cable 681 3629 995 Both
P260 and Cable 681 3630 1,995 Both
Serial Port Converter 61D 3925 25 Both
Cable, 9-Pin to 25-Pin
Asynch Printer/Terminal 61D 3926 50 Both
Cable, 9-pin to 25-pin,
4M
128 MB Level 3 Cache
Cache
4 x 32 MB, 400 MHz 681 4138 25,000 Both
5 x 32 MB, 433 MHz 681 4139 25,000 Both
8 MB Memory Card
Inward Facing 681 4181 65,536 Both
Outward Facing 681 4182 65,536 Both
16 GB Memory Card
Inward Facing 681 4183 114,688 Both
Outward Facing 681 4184 114,688 Both
32 GB Memory Card
Inward Facing 681 4188 163,840 Both
Outward Facing 681 4189 163,840 Both
4 GB Memory Card
Inward Facing 681 4196 36,864 Both
Outward Facing 681 4197 36,864 Both
Drawer Placement Indicator
Expansion Rack
1U Position 61D 4401 N/C Both
5U Position 61D 4405 N/C Both
9U Position 61D 4409 N/C Both
13U Position 61D 4413 N/C Both
19U Position 61D 4419 N/C Both
PCI Blind Swap Cassette 61D 4599 150 MES
Kit, Single Wide Adapters,
Universal
Drawer Placement Indicator
Primary Rack
1U Position 61D 4601 N/C Both
5U Position 61D 4605 N/C Both
9U Position 61D 4609 N/C Both
13U Position 61D 4613 N/C Both
18U Position 681 4618 N/C Both
Rack Indicator, Rack #1 61D 4651 N/C Both
61R Both
681 Both
IBM 64-bit/66 MHz PCI ATM 61D 4953 1,495 Both
155 UTP Adapter
Initial/
Model Feature Purchase MES/
Description Number Number Price Both
IBM 64-bit/66 MHz PCI ATM 61D 4957 $1,595 Both
155 MMF Adapter
IBM Token-Ring PCI Adapter 61D 4959 795 Both
IBM e-business Cryptographic 61D 4960 6,200 Both
Accelerator
10/100 Mbps Ethernet PCI 61D 4962 275 Both
Adapter II
PCI Cryptographic 61D 4963 7,000 Both
Coprocessor (FIPS-4)
Software Preinstall 681 5005 0 Initial
Preinstall
Initial/
Model Feature Purchase MES/
Description Number Number Price MMMC Both
4-Way POWER4 681 5240 $140,000 $2,392 Initial
HPC Processor
Option (First)
8-Way POWER4
Processor (First) 681 5242 200,000 3,417 Initial
Turbo Processor 681 5244 275,000 4,784 Initial
Option (First)
Turbo Processor 681 5245 275,000 4,784 MES
Option (First)
MES Feature
4-Way POWER4 HPC 681 5250 140,000 2,392 Both
Processor Option
Initial/
Model Feature Purchase MES/
Description Number Number Price Both
Processor Clock Card, 681 5251 $350 Both
Programmable
Initial/
Model Feature Purchase MES/
Description Number Number Price MMMC Both
8-Way POWER4 681 5252 $200,000 $3,417 Both
Processor Turbo 681 5254 275,000 4,784 Both
Option
Processor Bus Pass 681 5257 4,400 Both
Through Module
Initial/
Model Feature Purchase MES/
Description Number Number Price Both
Front Door
Black with Copper 61R 6070 $3,000 Initial
Accent, Primary Rack
Sculptured Black, 61R 6071 550 Both
Secondary Rack
OEM Rack, Primary (Black) 61R 6072 700 Initial
OEM Rack, Secondary 61R 6073 700 Both
(Black)
Rear Door, Primary or
Secondary Rack
Slim Line 61R 6074 750 Both
Acoustic 61R 6075 2,100 Both
I/O Drawer Attachment
Cable Group Drawer
Position
#4609 61D 6121 400 Both
#4605 61D 6122 400 Both
#4601 61D 6123 750 Both
#4613 61D 6124 750 Both
Initial/
Model Feature Purchase MES/
Description Number Number Price Both
#4401 61D 6125 $ 750 Both
#4405 61D 6126 750 Both
#4409 61D 6127 750 Both
#4413 61D 6128 750 Both
#4419 61D 6129 750 Both
20/40 GB 4 mm Internal 681 6158 2,995 Both
Tape Drive
Cable Group, Power 681 6161 400 Both
Controller to CEC and
Fans
Interface Cable, Service 681 6162 200 Both
Processor to Power
Subsystem
Power Converter Assembly
CEC 681 6170 4,700 Initial
I/O Drawer 61D 6172 4,000 Both
I/O Drawer to Media 61D 6179 300 Both
Drawer
Power Cable Group, CEC to
Power Controller
First Processor Module 681 6181 270 Both
Second Processor Module 681 6182 150 Both
Third Processor Module 681 6183 150 Both
Fourth Processor Module 681 6184 150 Both
Bulk Power Regulator 61R 6186 4,000 Both
Power Controller, Four 61R 6187 1,900 Initial
Cooling Fans, Three DC
Power Converter
Connections
Power Distribution 61R 6188 3,500 Both
Assembly, Ten DC Power
Converter Connectors
DC Power Converter, CEC, 681 6189 4,200 Both
Additional
Capacitor Book, Two 681 6198 1,800 Both
Processor Modules
Integrated Battery Backup, 61R 6200 4,500 Both
Primary
Integrated Battery Backup, 61R 6201 8,000 Both
Redundant
PCI Dual Channel Ultra3 61D 6203 790 Both
SCSI Adapter
PCI Universal Differential 61D 6204 650 Both
Ultra SCSI Adapter
PCI Single-Ended Ultra 61D 6206 395 Both
SCSI Adapter
Gigabit Fibre Channel 61D 6228 3,100 Both
Adapter for 64-bit PCI
Bus
Advanced SerialRAID Plus 61D 6230 3,000 Both
Adapter
128 MB DRAM Option Card 61D 6231 960 Both
32 MB Fast-Write Cache 61D 6235 575 Both
Option Card
Cable Group, Integrated 61R 6240 100 Both
Battery Backup to Power
Controller
Support Processor with 681 6404 7,000 Both
Remote I/O Loop
Attachment, Two Loops
Remote I/O Loop Adapter, 681 6410 8,000 Both
Four Loop
I/O Drawer PCI Planar, 10 61D 6563 8,000 Both
Slot, 2 Integrated Ultra3
SCSI Ports
Ultra3 SCSI 4-Pack Hot 61D 6564 500 Initial
Swap Back Plane
Backplane, CEC 681 6565 50,000 Both
Initial/
Model Feature Purchase MES/
Description Number Number Price Both
Quiet Touch Keyboard
U.S. English, #103P 681 6600 $ 100 Both
French, #189 681 6601 100 Both
Italian, #142 681 6602 100 Both
German/Austrian, #129 681 6603 100 Both
U.K. English, #166 681 6604 100 Both
Spanish, #172 681 6605 100 Both
Japanese, #194 681 6606 100 Both
L.A. Spanish, #171 681 6629 100 Both
Russian, #443 681 6638 100 Both
U.S. English ISO9995, 681 6640 100 Both
#103P
Hardware Management 681 7315 4,325 Both
Console I Attachment
Cable, Hardware
Management Console
to Host,
6-Meter 681 8120 50 Both
15-Meter 681 8121 75 Both
128-Port Asynchronous
Controller Cable
4.5 Meter 61D 8131 60 Both
23cm (9-Inch) 61D 8132 40 Both
RJ-45 to DB-25 Converter 61D 8133 120 Both
Cable
Enhanced Remote 61D 8137 1,595 Both
Asynchronous Node
16-Port EIA-232
Line Cord
6AWG, 14ft, No Plug 61R 8677 500 Both
6AWG, 14ft, IEC309 Plug 61R 8678 1,000 Both
10AWG, 14ft, IEC309 Plug 61R 8680 750 Both
6AWG, 6ft, Chicago 61R 8681 750 Both
IEC309 Plug
10AWG, 6ft, Chicago 61R 8682 750 Both
IEC309 Plug
Bulk Power Assembly, 61R 8690 5,000 Initial
Redundant
Expansion Rack, 24-Inch, 61R 8691 8,000 Both
42U
Media Drawer, Operator 681 8692 1,700 Both
Panel, Diskette
Quiet Touch Keyboard --
USB, Stealth Black, U.S.
English, #103P 61D 8800 100 Both
English (EMEA), #103P 61D 8840 100 Both
Mouse -- Stealth Black 61D 8841 75 Both
with Keyboard Attachment
Cable
Language Group Specify
U.S. English 61D 9300 N/C Initial
61R Initial
681 Initial
French 61D 9703 N/C Initial
61R Initial
681 Initial
German 61D 9704 N/C Initial
61R Initial
681 Initial
Polish 61D 9705 N/C Initial
61R Initial
681 Initial
Portuguese 61D 9707 N/C Initial
61R Initial
681 Initial
Spanish 61D 9708 N/C Initial
61R Initial
681 Initial
Initial/
Model Feature Purchase MES/
Description Number Number Price Both
Italian 61D 9711 N/C Initial
61R Initial
681 Initial
Canadian French 61D 9712 N/C Initial
61R Initial
681 Initial
Japanese 61D 9714 N/C Initial
61R Initial
681 Initial
Korean 61D 9716 N/C Initial
61R Initial
681 Initial
Hungarian 61D 9719 N/C Initial
61R Initial
681 Initial
Slovakian 61D 9720 N/C Initial
61R Initial
681 Initial
Russian 61D 9721 N/C Initial
61R Initial
681 Initial
Simplified Chinese (PCR) 61D 9722 N/C Initial
61R Initial
681 Initial
Power Cord Specify
U.S./Canada 61D 9800 $0 Both
681 Both
Belgium, Finland, France 61D 9820 0 Both
(250V, 16A) 681 Both
Denmark (250V, 10A) 61D 9821 0 Both
681 Both
U.K. (250V, 13A) 61D 9825 0 Both
681 Both
Israel (250V, 6-16A) 61D 9827 0 Both
681 Both
Switzerland (250V, 10A) 61D 9828 0 Both
681 Both
South Africa/Pakistan 61D 9829 0 Both
(250V, 16A) 681 Both
Italy (250V, 10A and 16A) 61D 9830 0 Both
681 Both
Argentina (250V, 10A) 61D 9831 0 Both
681 Both
Thailand (250V, 15A) 61D 9833 0 Both
681 Both
Uruguay (250V, 10A) 61D 9834 0 Both
681 Both
For ServiceElect (ESA) Maintenance Service Charges, contact IBM Global
Services at 888-IBM-4343 (426-4343).
Feature Conversion Purchase Price
Feature
Conversion
Purchase
From To Price(2)
5240 - 4-Way, POWER4, 5245 - 8-Way, POWER4, $240,000
HPC Turbo
5250 - 4-Way, POWER4, 5254 - 8-Way, POWER4, 240,000
HPC Turbo
5242 - 8-Way, POWER4 5245 - 8-Way, POWER4, 225,000
Turbo
5252 - 8-Way, POWER4 5254 - 8-Way, POWER4, 225,000
Turbo
4138 - 128 MB L3 Cache, 4139 - 128 MB L3 Cache, 18,750
400 MHz 433 MHz
- 2
- If field installed on a purchased machine, parts removed or replaced become the property of IBM and must be returned.
Continuous Maintenance: Yes for processors; No for L3 cache
Customer Financing: IBM Global Financing offers attractive financing to credit-qualified commercial and government customers and Business Partners in more than 40 countries. IBM Global Financing is provided by the IBM Credit Corporation in the United States. Offerings, rates, terms, and availability may vary by country. Contact your local IBM Global Financing organization. Country organizations are listed on the Web:
Order Now
Use Priority/Reference Code: RE001
Phone: 800-IBM-CALL
Fax: 800-2IBM-FAX
Internet: ibm_direct@vnet.ibm.com
Mail: IBM Atlanta Sales Center
Dept. RE001
P.O. Box 2690
Atlanta, GA 30301-2690
You can also contact your local IBM Business Partner or IBM representative. To identify them, call 800-IBM-4YOU.
Note: Shipments will begin after the planned availability date.
Trademarks