White Paper Issue January 27 2001

Storage systems for the high-end enterprise computing

The Symmetrix 8730 system is a powerful disk storage system that is suitable for use with BS2000/OSD and all open systems marketed by Fujitsu Siemens Computers, as well as in networked environments with open systems from other vendors. The system combines high performance with virtually 100% fault tolerance, which makes it the ideal choice for meeting current and future requirements for business-critical storage solutions. The Symmetrix 8730 system extends the successful Symmetrix product family in the top capacity class and perfectly complements the storage systems from Fujitsu Siemens Computers.

The Symmetrix 8730 system includes two model types: Symmetrix 8730-36 and Symmetrix 8730-73, providing storage capacities up to 11.7 Terabytes (RAID-1-formatted for open systems). It comes equipped with high-performance disk drives with nominal capacities of 36 and 73 Gbytes.

The Symmetrix 8730 system possesses all of the functions and features of the Symmetrix disk storage system family. Armed with adequate performance and capacity reserves, this system forms the basis for powerful enterprise storage solutions for high-end systems.

The Symmetrix 8730 disk storage system can be equipped with SCSI, Fibre Channels and ESCON system connections, and uses a cache memory which is protected against power failure and other faults. The cache is scalable up to 32 Gbytes, ensuring very good performance values thanks to high hit rates when reading data and use of the “store-in” principle for writes. A choice of RAID configurations (RAID-1 or RAID-S) and the ability to mirror data locally or remotely on separate subsystems, even over considerable distances and independent of the connected server systems, are a guarantee of optimum adaptability to the user’s security and availability requirements. With the optional multi-mirroring feature TimeFinder, parallel use of data resources by independent applications is possible.

Special features

• Efficient, centralized storage environments for BS2000 and open systems servers from Fujitsu Siemens Computers or other open systems vendors. Up to 64 host connections are available.
• A high level of integration combined with innovative drive technology ensures maximum capacity, availability and cost-effectiveness. Up to 384 disk drives may be built in.
• RAID-S and RAID-1 (local and remote) deliver optimum performance and availability.
• Designed to provide uninterrupted operation and service, allowing continuous production and online upgrades.
• A uniform architecture for all Symmetrix systems guarantees continuity of familiar functional features and enhancement products.
  

System environments

The Symmetrix 8730 storage system integrates perfectly with BS2000 and open systems from Fujitsu Siemens Computers, but is also compatible with other open system platforms. It is supported by BS2000 systems, PRIMEPOWER servers running Solaris, RM servers running Reliant Unix, and the Intel-based PRIMERGY servers running Windows NT, Windows 2000 and Linux. Optimal availability configurations can be realized by integrating Symmetrix systems, including add-on products such as RMS (Reliant Monitor Software) with the Reliant UNIX or Solaris operating systems, or MSCS (Microsoft Cluster Server) with Windows NT systems. Equipped with extension functions for optimized performance, data and operational reliability, and application availability, the Symmetrix systems meet the requirements of standardized and centralized data storage management.

Technology

The storage systems in the Symmetrix 8000 system family use the same architecture across all model variants and therefore offer largely the same functionality, performance and availability in all configurations. The main differences are in their expandability in terms of disk capacity, host connectivity and cache size. Every system consists of the following main components:

• Basic system with infrastructure, redundantly configured components such as system bus, power supply, AC power distributor and fans, and the service processor with modem for remote maintenance
• Channel Directors with Ultra SCSI, Fibre Channel and/or ESCON interfaces for connection to servers
• Disk Directors for parallel access to all drives
• Cache memory with battery backup to protect against power failure
• Disk modules with high-performance 36 and 73 GB 10,000 rpm hard disks

The operational reliability and availability of Symmetrix systems are guaranteed thanks to RAID-1 and RAID-S technology and the capability to duplicate data directly either locally or on a second Symmetrix system set up at a remote location (SRDF). Availability is enhanced further by full hardware redundancy and online replacement of components without interruption to system operation.

Reliability functions

• Data mirroring

Symmetrix systems satisfy all availability levels, especially also the highest level, DTDS+ (Disaster Tolerant Disk System plus), as defined in the original RAID standards of the University of Berkeley, as well as in the new classifications specified by the RAID Advisory Board (RAB). Users can combine data mirroring with RAID-1 and multi-mirroring (TimeFinder) either locally or remotely. This allows an optimum tradeoff between performance, availability and costs.

• Disaster Protection

The SRDF function (Symmetrix Remote Data Facility) provides protection against data loss in disaster scenarios by maintaining a copy of the data resource on a second remote system, without placing any additional load on the host. If the first system fails, processing can continue with the mirrored data resource on the other system. The systems are connected by redundantly con-figured high-speed fiber optic links, each supporting data transfer rates up to 100 Mbyte/s over distances up to 60 km (with repeaters). There is also a special operating mode in which data can be mirrored over virtually unlimi-ted distances. The SRDF option is available for all Symmetrix models, and can also be used between different, or differently configured, models. As well as the hardware, the appropriate SRDF function firmware must be installed in each of the two systems.

• Failure protection

Potential problems in the cache and drives are usually detected early by regular self-test functions running in the background. If the cache is affected, the corresponding cache areas are detached. If drive faults are detected, the data is copied from the active drive to the most suitable standby spare drive without disruption to online system operation. Necessary maintenance activities can subsequently be performed without downtime.

Cache memory management

Cache memory is one of the main components in Symmetrix systems. It allows input/output operations to be performed at electronic speed while simultaneously acting as a data buffer for device emulation. This results in logical volumes being separated both physically and in time from individual disk drives. The cache is connected to the other system components by a powerful, redundant bus system and is protected by a battery buffer against data loss in the event of power failure.

• Cache algorithms

Efficient use of cache capacity is ensured by cache algorithms which react dynamically to the current data traffic depending on access patterns. Thus, sequential access operations are handled differently from randomly distributed input/output operations. Appropriate host instructions can also be implemented in individual cases.

• Cache protection mechanisms

The cache is protected against all individual component failures by redundant components in the disk configuration and ECC/ERC protection on all data paths. Defective memory components are isolated and taken offline in the event of failure. If there is a power failure, system operation is maintained with the aid of the battery buffer for a sufficient time to enable the entire contents of the write cache to be written to free disk space.

Remote service support

In order to ensure a rapid, targeted response to new or existing problems, each subsystem has a direct modem connection to the Symmetrix service center. The service center identifies, coordinates and initiates all the necessary actions. If a problem occurs or internal thresholds are exceeded (e.g. number of read or write retries on the hard disk), an auto-call is automatically made – subject to the customer’s agreement – to the EMC² service center, where the problem is diagnosed and the required actions initiated.

Functional enhancements

• Enterprise Storage Platform

The ESP option opens up the centralized use of Symmetrix systems for BS2000/OSD and other system platforms from Fujitsu Siemens Computers. Supported operating systems include Reliant UNIX from Fujitsu Siemens Computers, Solaris, and Windows NT, as well as systems from other vendors. ESP brings considerable improvements in terms of cost-effective administration, data backup and shared use of data resources in enterprise computing. Data is stored on different physical drives so that a drive crash has no effect on the data belonging to other system platforms.

• Symmetrix Remote Data Facility (SRDF)

To support data mirroring on a remote Symmetrix system, the local and remote Symmetrix devices each require the appropriate function firmware. In synchronous mode, the systems are physically connected over distances up to 60 km (with repeaters) using redundantly configured Remote Link Directors and fiber optic links. With asynchronous data mirroring, distances of several hundred kilometers are possible.

• TimeFinder

Steadily growing volumes of data and the increasing load on computer systems are gradually reducing available time windows for necessary activities such as data backup, test runs or data mining. The TimeFinder option is available for all Symmetrix systems. TimeFinder generates one or more synchronous mirrored data sets (BCVs) which, when taken offline, can be used for the specified purposes. Resynchronization of the data can be performed automatically by TimeFinder following job completion.
The mirrored data sets (Business Continuance Volumes) generated by TimeFinder can be used for test and backup purposes in parallel with the real production data. When integrated with the NetWorker data backup tool, BCVs can be used in online mode, for example, to back up ORACLE databases without any adverse effect on online operation.

• Volume Logix

The Volume Logix function defines a dedicated relationship between the logical volumes in the Symmetrix system and the connected servers in configurations in which multiple servers are connected via Fibre Channel hubs or switches with one Symmetrix port. In heterogeneous system environments, Volume Logix centrally controls access to the Symmetrix system.

• Control Center

The ControlCenter Manager is an optional software product allowing centralized monitoring of both the operational status and the capacity utilization of a Symmetrix system. It is operated via the graphical user interface (Windows GUI) of a workstation or cluster console. In addition to displaying the configuration, it can generate information on logical volume sizes, protection mechanisms and performance statistics. This information allows data throughput bottlenecks to be identified and aids early implementation of specific tuning measures.
The Control Option enhances CC SRDF/TimeFinder configurations with suitable functionality, e.g. for displaying the status and utilization level of remote links, including the synchronization status of data mirroring. In addition, scripts included in the software allow automation of monitoring and recovery routines.
The CC Disk Reallocation option allows the operator to assign logical volumes to channels or servers, thus optimizing performance and simplifying resource allocation.
The CC Workload Analyzer option is additionally provided to optimize performance analysis and tuning measures.
The ControlCenter Package includes the basic license and the three options described above.

• PowerPath

The PowerPath function is a server-resident software solution for dynamically distributing read and write I/O operations over max. 32 SCSI or 16 Fibre Channel ports. Distribution of I/O operations over all connected channels can be implemented according to different criteria (e.g. on the basis of the number of remaining I/O operations, the number and size of blocks to be transferred, Symmetrix optimization rules, etc.). If a channel fails, it is taken offline to detach it from the configuration and the remaining I/O operations are distributed over the remaining channels. After replacement of the defective channel, a recovery procedure is initiated either automatically or manually. PowerPath provides functions for optimizing data transfer between Symmetrix systems and connected servers without the overhead of laborious reconfiguration. Even in the event of a system adapter failure, continuity of operation is guaranteed. PowerPath is available for MS Windows NT platforms such as the PRIMERGY series, for Solaris on PRIMEPOWER (GP7000F) systems and Reliant UNIX on RM600E systems from Fujitsu Siemens Computers.

Product overview

System cabinets

Disk storage units (factory or field installation)

Cache memory

System adapters

Channel expansions

System licenses

B5TF-OPEN-CTxx

D:B5VLGX-CTxx

xx = 1 to 12 (tiers 1 to 12 depending on raw capacity of Symmetrix 8730 system)

Technical data

Symmetrix 8730 disk storage system (specifications are based on maximum configuration)

36Gbyte disks

73 Gbyte disks

36 Gbyte

73 Gbyte

Electrical specifications

Environmental specifications

Mechanical specifications

Compliance with standards

Note:The following applies to all data volumes quoted (according to DIN 1301 and DIN 44300):
1 kbyte = 10³ bytes, 1 MB = 106 bytes, 1 GB = 109 bytes, 1 TB = 1012 bytes.
(1) RAID-S is not supported on 73 Gbyte disk drives.
(2) Limited to 320 drives if 36 and 73 Gbyte types are used in mixed configurations.