A Windows VPS arrives with a virtual network adapter attached to the provider’s infrastructure, and by default that adapter pulls its address from DHCP. For a server that runs IIS, SQL Server, or RDP, that default works, but a deliberately configured network stack \u2014 a stable IP, predictable DNS servers, and a verified default route \u2014 makes administration safer, debugging faster, and outage recovery simpler.<\/p>\n\n\n\n
Network behavior differs noticeably between providers: some hand out a single vNIC with one IP, others support multiple vNICs, extra IPv6 blocks, or bandwidth caps on the virtual switch. Compare Windows VPS providers on our comparison table<\/a> before you commit, so the configuration steps below match what your host actually offers.<\/p>\n\n\n\n Before changing anything, capture the current state. On Windows Server 2019 and later, the Write down the interface alias (usually With the values captured, assign the static address and DNS servers. The provider’s control panel normally shows the exact IP, subnet mask, and gateway allocated to the instance; use those numbers verbatim:<\/p>\n\n\n\n You can achieve the same result through Network Settings > Change adapter options<\/strong> in Server Manager, but the cmdlets are reproducible and safe to run inside a script. The critical operational rule: never close your RDP session until you have verified connectivity from a second session<\/strong>. A wrong gateway or prefix length drops the connection instantly, and if the provider’s console offers no out-of-band access you can lock yourself out entirely.<\/p>\n\n\n\n After applying the settings, confirm every layer of the stack \u2014 interface, route, DNS, and reachability:<\/p>\n\n\n\n Many providers assign an IPv6 prefix alongside IPv4, and Windows enables IPv6 by default. That is usually fine, but it creates two failure modes: applications that bind only to If you keep IPv6, configure the DNS servers for it explicitly ( NIC teaming combines multiple physical adapters into one logical interface for redundancy and throughput. On a VPS this only makes sense if the provider exposes more than one virtual NIC backed by distinct queues or bandwidth allocations \u2014 otherwise you are teaming two virtual adapters that share the same physical uplink, which adds complexity without adding capacity. Check your plan: single-vNIC plans (the majority) should skip teaming entirely.<\/p>\n\n\n\n On a provider that supports two vNICs, create the team with:<\/p>\n\n\n\n Apply the static IP to the team interface afterward, and test failover by disabling one member adapter while a ping runs. If the provider’s virtual switch does not support LACP, Network changes on a remote server should always have a rollback path. Keep the original DHCP values in a comment block at the top of your configuration script, and know the one-liner that returns the adapter to DHCP:<\/p>\n\n\n\n Schedule risky changes in a maintenance window, keep a second RDP session or the provider’s web console open as a lifeline, and test from a different network (for example, a phone on LTE) so you are not testing connectivity from the very path you might have broken.<\/p>\n\n\n\n A deliberate network configuration \u2014 static IP, explicit DNS, verified routes, and (only where the provider supports it) NIC teaming \u2014 removes a whole class of intermittent failures from Windows VPS administration. Capture the current state, apply changes with the cmdlets above, verify every layer before closing your session, and keep the DHCP rollback one-liner handy. To size a host with the vNIC layout and bandwidth your plan needs, see the full specs and pricing<\/a> on the Windows VPS comparison table.<\/p>\n\n\n\nInspect the Current Network Configuration<\/h2>\n\n\n\n
NetTCPIP<\/code> module gives you the full picture in four commands:<\/p>\n\n\n\nGet-NetIPConfiguration\nGet-NetIPAddress -AddressFamily IPv4 | Format-Table InterfaceAlias, IPAddress, PrefixLength\nGet-DnsClientServerAddress -AddressFamily IPv4\nGet-NetAdapter | Format-Table Name, InterfaceDescription, Status, LinkSpeed<\/code><\/pre>\n\n\n\nEthernet<\/code> or Ethernet Instance 0<\/code>), the current IP, the gateway, and the DNS servers. On a provider network, the gateway is almost always the first usable address in the subnet, and the DNS servers are the provider’s resolvers \u2014 do not invent your own gateway or you will lose connectivity the moment you apply the static configuration.<\/p>\n\n\n\nSet a Static IP Address with PowerShell<\/h2>\n\n\n\n
New-NetIPAddress -InterfaceAlias \"Ethernet\" `\n -IPAddress 203.0.113.10 -PrefixLength 24 -DefaultGateway 203.0.113.1\nSet-DnsClientServerAddress -InterfaceAlias \"Ethernet\" `\n -ServerAddresses (\"1.1.1.1\",\"8.8.8.8\")<\/code><\/pre>\n\n\n\nVerify DNS Resolution and Connectivity<\/h2>\n\n\n\n
Test-NetConnection 203.0.113.1 -Port 3389 # gateway reachable?\nResolve-DnsName windows-vps.org # DNS resolves?\nGet-NetRoute -DestinationPrefix \"0.0.0.0\/0\" # default route present?<\/code><\/pre>\n\n\n\nTest-NetConnection<\/code> doubles as a firewall probe, which makes it useful for verifying that RDP (3389) and your web bindings are reachable after a network change. If resolution fails, check the DNS server list first \u2014 a stray 127.0.0.1 entry from a local DNS service is a common cause of “name not resolved” on freshly configured VPS instances.<\/p>\n\n\n\nIPv6 Considerations on a VPS<\/h2>\n\n\n\n
::<\/code> may answer on IPv6 while your monitoring checks IPv4, and some older RDP clients prefer the IPv4 address when both exist. If you do not use IPv6, disable it cleanly rather than leaving a half-configured stack:<\/p>\n\n\n\nGet-NetIPAddress -AddressFamily IPv6 | Format-Table InterfaceAlias, IPAddress\nDisable-NetAdapterBinding -Name \"Ethernet\" -ComponentID ms_tcpip6<\/code><\/pre>\n\n\n\nSet-DnsClientServerAddress -AddressFamily IPv6<\/code>) and add firewall rules for the IPv6 addresses you actually use. A VPS with a misconfigured IPv6 stack will intermittently fail connections that resolve to AAAA records.<\/p>\n\n\n\nNIC Teaming on a VPS: When It Helps and When It Does Not<\/h2>\n\n\n\n
New-NetLbfoTeam -Name \"Team1\" `\n -TeamMembers \"Ethernet\",\"Ethernet 2\" `\n -TeamingMode SwitchIndependent `\n -LoadBalancingAlgorithm Dynamic\nGet-NetLbfoTeam | Format-List Name, Status<\/code><\/pre>\n\n\n\nSwitchIndependent<\/code> mode is the only safe choice \u2014 static teaming against a switch that expects LACP will black-hole the whole interface.<\/p>\n\n\n\nApply Changes Safely and Roll Back<\/h2>\n\n\n\n
Set-NetIPInterface -InterfaceAlias \"Ethernet\" -Dhcp Enabled\nSet-DnsClientServerAddress -InterfaceAlias \"Ethernet\" -ResetServerAddresses<\/code><\/pre>\n\n\n\nSummary<\/h2>\n\n\n\n