Network configuration under SERVERware
Non-redundant network
The diagram above shows the typical configuration of SERVERware 4.
The system is built around storage hosts, processing hosts, and backup hosts. Two storage hosts configured to work as mirrors provide storage for the processing hosts on SAN (Storage Array Network). The processing hosts run locally on the VPSs provided on the storage network. The VPSs use the WAN network to provide services to the customer.
The backup host provides disk space for backing up the VPSs. Backup hosts use the storage network to transfer data from the storage hosts to the local backup disk space.
Due to the nature of storage traffic, we need to separate SAN traffic from other traffic on the network. In the case where we have many processing hosts, the best solution is to use a dedicated network switch that serves the network traffic. In reality, we will use the same switch for the SAN and any other traffic.
In this case, we need to separate the traffic by using the VLAN concept. Thus, for SAN traffic, a SAN VLAN will be created. Other traffic we can keep in the default native VLAN. Optionally, we can create VLANs for voice traffic and configure the priority of this traffic.
The following schematic diagram shows the SERVERware 4 network with a Cisco switch (no network redundancy):
To separate SAN traffic from other traffic on the network, we can use VLANs and configure all ports belonging to the SAN network to be members of the SAN VLAN.
Here is a Cisco configuration example that shows how to configure VLANs on the ports:
Redundant network
The SERVERware 4 network design with redundancy at the network level is based on two network switches connected and configured to operate in a stack.
The schematic diagram of SERVERware 4 network redundancy is shown in the following image:
As shown in the diagram above, the redundant network is based on two switches Cisco WS-3750G . The switches are connected in a stack using appropriate stacking cables. The WS-3750G stacking bus supports 32 Gbit/s. Since we have two switches, the hosts must be connected to both switches. This means that for each connection (SAN, WAN), the host must have two Ethernet interfaces connected to the two switches.
This network design allows for redundancy at the network level. In case of a switch failure, another switch will take over the network traffic.
The Cisco switch supports EtherChannel on switch stacks, which is necessary, and without it,
the network redundancy will not work.
In the same way that we ensure the separation of SAN traffic in the non-redundant network design, in the redundant network design as well, to separate SAN traffic from other traffic, we will configure the appropriate VLANs. Here is an example Cisco switch configuration with network redundancy and VLANs configured to separate SAN traffic:
