The architecture of the IBM® DS8000
is based on three major elements that provide function specialization
and three tiers of processing power.
Figure 1 illustrates the
following elements:
- Host adapters manage external I/O interfaces that use Fibre Channel
protocols for host-system attachment and for replicating data between
storage systems.
- Flash RAID adapters and device
adapters manage the internal storage devices. They also manage the
SAS paths to drives, RAID protection, and drive sparing.
- A pair of high-performance redundant active-active Power® servers is functionally positioned between
the adapters and a key feature of the architecture.
The internal Power servers support the bulk
of the processing to be done in the storage system. Each Power server has multiple processor cores.
The cores are managed as a symmetric multiprocessing (SMP) pool of
shared processing power to process the work that is done on the Power server. Each Power server runs an AIX® kernel that manages the processors, manages
processor memory as a data cache, and more. For more information,
see IBM DS8000 Architecture
and Implementation on the IBM Redbooks® website.
Figure 1. DS8000 architecture
The DS8000 architecture has the following major benefits.
- Server foundation
- Promotes high availability and high performance by using field-proven Power servers
- Reduces custom components and design complexity
- Positions the storage system to reap the benefits of server technology
advances
- Operating environment
- Promotes high availability and provides a high-quality base for
the storage system software through a field-proven AIX operating-system
kernel
- Provides an operating environment that is optimized for Power servers, including performance
and reliability, availability, and serviceability
- Provides shared processor (SMP) efficiency
- Reduces custom code and design complexity
- Uses Power firmware and
software support for networking and service functions