In the current challenging business environment, companies are looking to cut their deployment and operating costs by consolidating servers and storage into datacenters. They are also becoming increasingly interested in intra-corporate cloud computing and other IT infrastructure optimization strategies that link multiple systems seamlessly and efficiently using SOA.
In parallel, those same companies are being viewed by the public not only for the products and services they provide, but also on their business continuity, stability, legal compliance, and environmental actions.
In this climate, and as a result of the mainframe server's crucial role in centralized data processing, its importance is being re-acknowledged.
GlobalServer, with its enhanced performance and reliability, coupled with green IT credentials is able to contribute to customer business growth and improved customer value.
Improved processing performance
- Higher performance, higher reliability
GS21 2600 model group adopts a system-on-chip design with the latest CMOS technology to boost transaction-processing performance up to 40% over previous models, while also increasing reliability. Each chipset hosts up to 8 cores, equipped with a 256 KB primary cache, as well as a secondary cache of up to 24 MB, an I/O processor, a memory controller, and a system controller.
- Expanded scalability
When configuring large-scale virtual machines (VM) with around 20,000 virtual I/O devices as a development environment, previous models were limited to configurations of three VM per cluster. However, GS21 2600 models offer both hardware and software enhancements that enable a dramatically larger number of virtual I/O configurations of over 10 VM. This allows concurrent development, revision and testing for different applications regardless of scale.
Hardware feature for high-reliability
- Instruction retry function
On detection of a hardware error during instruction execution, the CPU automatically retries the instruction. This enables automatic recovery from intermittent hardware errors.
- Automatic fallback feature
The translation look-aside buffer (TLB) and buffer storage (BS) consist of several ways (divisions of the buffers) and have an automatic fallback feature. When an intermittent error occurs in the TLB or BS, recovery is performed by re-executing the instruction, But if a permanent error occurs, execution retry will be unsuccessful. In this case the faulty part is automatically isolated in way units, and processing is continued using the other available ways.
- Automatic alternate memory allocation feature
Main storage and system storage, where important data and instructions are held, has an automatic alternate memory allocation hardware feature for handling permanent errors on the memory device.
A memory patrol feature, which periodically reads and checks the data stored in main memory and system storage, helps prevent storage errors.
Green ICT approach
- Lower power consumption
GS21 2600 model group use system-on-chip technology to consolidate what had been 14 separate chipsets into a single chipset. Furthermore, features such as a compact, high-performance power unit reduce power consumption by as much as 50% over previous models, with the same processing performance.
- Higher performance power supply
GS21 2600 model group adopt a compact high-performance 80 PLUS GOLD power supply units , leading to stable power supply with a small energy loss.