ページを印刷する
Build 252-node POD/cluster with Flex System Interconnect Fabric
With the growth of cloud, media applications, mobile connections and big data clients, IT departments are faced with
many new requirements. Flex System Interconnect Fabric is designed to meet client needs by providing a simple
Ethernet fabric cluster that accelerates deployment, simplifies management, and enables dynamic scalability,
increases reliability, availability and security in medium to large scale POD deployments. This solution offers a solid
foundation of compute, network, storage, and software resources in a Flex System POD.
The key I/O components of this solution consist of a pair of RackSwitch G8264CS. One of the G8264CS will be the
center of intelligence and provides all direction and updates to the redundant G8264CS and the 2-18 Flex System
SI4093 System Interconnect Modules. By using Flex System x222 compute nodes, clients can easily setup a single
chassis and then scale up to nine chassis with easy to build a 252-node POD or cluster. In addition, the automated
capabilities of adding additional chassis after initial setup of the first client can also exploit the acquisition and
operation cost savings of converging Ethernet and Fibre Channel traffic within the POD or cluster, but still be able to
simply connect into their existing upstream networks.
Figure 10 shows the Flex System Interconnect Fabric using 9 chassis with SI4093, and a pair of RackSwitch
G8264CS connecting to a clients existing LAN and SAN which could be a Brocade switch or a Cisco MDS switch.
Figure 10. Flex System Interconnect Fabric using 9 chassis
The solution components that are used in the scenarios that are shown in Figure 10 are listed in Table 11.
Table 11. Building a Flex System Interconnect Fabric POD using FCoE (Figure 10)
RackSwitch G8264CS
23