Skip to main content

Worksheet considerations

This section covers the considerations required for filling out the pre-install network configuration worksheet.

Take into account the following when filling out the worksheet:

  • Data Center Uplink

    The data center uplink is the main data path of all northbound and southbound traffic in support of the ThinkAgile CP environment. This VLAN is customer defined and lives on the customer network. This network needs connectivity to the Internet (see Firewall Settings). It also needs a unique IP address and subnet across the ThinkAgile CP environment.

    • For dual interconnect configurations, you must provide one IP address for each interconnect switch and one IP address for the Virtual Router Redundancy Protocol (VRRP).

    • For single interconnect configurations, only one IP address is required.

  • OOB Interconnect network

    This network is used to manage the interconnect switches; it provides hardware health reporting and management between the interconnect switches and the cloud controller. This network needs connectivity to the Internet (see Firewall Settings). It also needs a unique IP address and netmask across the ThinkAgile CP environment.

  • Hypervisor network

    The hypervisor network is used to provide internal connectivity among the host operating systems of the compute and storage controllers, and the virtual networking (VNET) NFV appliances. This network must be unique and routable from the customer network.

  • Hardware management network

    This network connects the System Management Module (SMM) for each compute enclosure, the Cloud Controller for each compute node, and the Baseboard Management Controller (BMC) for each storage controller.

  • Multi-enclosure link aggregation network (MLAG)

    This network operates on the Interconnect Switch Link (ISL) between the two interconnects and is used for internal control communication associated to the multi-enclosure link aggregation protocol between the interconnect switches. This network only requires two IP addresses, and by default, the system uses addresses taken from a reserved range.

  • Variable names

    Variable names are shown in italics in worksheet tables. Those names correspond to parameters that you need to specify when deploying the ThinkAgile CP solution (using the ThinkAgile CP Deployment Guide).

Prior to installation, use the following table as a guide to the configuration details required for a successful ThinkAgile CP deployment.

Table 1. Configuration details guidelines

Network

IP Address Requirements

Assignment Details

Address Type

Data Center Uplink

1 IP address per Interconnect

1 Virtual IP Address (VRRP)

Statically configured at installation time

public1

OOB Interconnect

1 IP address per Interconnect

Statically configured at installation time

public

Hardware management network VLAN2

1 IP address per Interconnect

Statically configured at installation time

private

Hardware management network IP Range3

1 IP address per compute node's XCC

1 IP address per compute block SMM

1 IP address per storage controller

Dynamically configured by the DHCP server on the interconnects

private

Hypervisor VLAN

1 IP address per Interconnect

1 Virtual IP Address (VRRP)

Statically configured at installation time

public

Hypervisor IP Range

1 IP address per compute node

4 IP addresses per storage controller

Remaining IP addresses that can be assigned to application VMs vNICs and outside interfaces of NFV VMs only when the corresponding vNICs are configured in VLAN mode and assigned to the default VLAN

Dynamically configured by the DHCP server on the interconnects

public

MLAG Peers

1 IP address per Interconnect

Hard-coded values, which may be optionally be statically configured at installation time

private

VNET IP Ranges

Variable number of IP addresses, which are assigned application VM’s vNICs configured in VNET mode

Dynamically configured by the DHCP server on the NFV VM

public or private

VNET Outside Interface (optional)

1 IP address per NFV, assigned to the outside Interface of an NFV VM configured in VLAN mode and assigned a non-default VLAN

Dynamically configured by a DHCP server on the customer data center network or statically at VNET creation time

public

Application VM’s vNICs (optional)

1 IP address per vNIC, assigned to an application VM’s vNIC configured in VLAN mode and assigned a non-default VLAN

Dynamically configured by a DHCP server on the customer data center network

public

1 The term public means routable in the customer data center network. In other words, any host connected to this network must be able to reach these addresses.
2 These IP addresses are public in development environments and labs, as to allow access for debugging purposes and operating system resets during development. However, for customer installations, the recommendation is to use private IP addresses, which must not be accessible.
3 These IP addresses are public in development environments and labs, as to allow access for debugging purposes and operating system resets during development. However, for customer installations, the recommendation is to use private IP addresses, which must not be accessible.