The Network section on serverplus: networks, routers, IPs and balancers

The "Network" section in the cloud.serverplus.uz panel is everything related to your cloud's network infrastructure: how servers are connected to each other, how they reach the internet and how you protect and distribute traffic. The section is split into tabs, and in this article we go over what each one is responsible for.

The section's tabs: Networks, Subnets, Routers, Floating IP addresses, Ports, Load balancers, Security groups and VPNs.

How it all connects

The easiest way is to picture the network stack from the bottom up:

• A network is a virtual "switch" that servers connect to. Inside it lives a subnet — a range of IP addresses.
• A server connects to the network through a port (a virtual network card) and gets an internal IP from the subnet.
• A router connects your private network to the external one and provides internet access.
• A floating IP is a public address you "hang" on a server so it is reachable from outside.
• A security group filters traffic (which ports are open), while a load balancer distributes the load across several servers.
• A VPN connects the cloud to your office or another site over a secure channel.

Networks

A network is an isolated virtual segment, the equivalent of a physical switch. Servers on the same network "see" each other directly. When creating one (+ Create network), you set:

• The network's name and description.
• A subnet — usually created together with the network right away (see below): CIDR, gateway, DHCP, address pool.

Each network has parameters: Status (ACTIVE), External router (whether the network is external), Shared, MTU (usually 1442). Private networks are free (0 UZS).

Subnets

A subnet is a range of IP addresses inside a network. It is exactly from the subnet that servers get their internal addresses. Key parameters:

• CIDR — the range, for example 10.0.0.0/24 (256 addresses).
• Gateway IP address — usually the first address, for example 10.0.0.1.
• DHCP — automatic address handout to servers (on by default).
• IP version — IPv4 or IPv6.
• Allocation pools — from which range to hand out addresses (for example from 10.0.0.10 to 10.0.0.200).

Routers

A router connects your private network to the external network (the internet) and routes traffic between subnets. A router has an external gateway — the public address through which the private network reaches outside.

• What the list shows: name, status (ACTIVE), gateway (external IP), region.
• Why it is needed: without a router, servers in a private network are isolated and have no internet access. It also lets different subnets talk to each other.

Floating IP addresses

A floating IP is a public address that can be assigned to a server and removed from it at any moment. The server lives with an internal IP, while the "floating" one makes it reachable from the internet.

• Status DOWN — the address is reserved but not attached to anything; ACTIVE — attached to a server.
• Fixed IP address — the server's internal address (10.0.0.186) the floating one "points" to.
• Convenience: you can move the same public address from server to server (for example, when replacing a machine) — the address does not change for external users.

Ports

A port is a virtual network card on a network. Every connection to a network is a port with its own IP and MAC address. Ports are created automatically, but they are visible and can be created manually.

• Device owner shows who the port belongs to: compute:nova — a server's network card, network:router_interface — a router interface, network:distributed — a service one (DHCP).
• Status ACTIVE/DOWN — whether the port is in use.

In most cases you do not need to touch ports manually — they appear on their own when you create servers, routers and load balancers.

Load balancers

A load balancer distributes incoming traffic across several servers — to handle load and not lose availability if one server fails. When creating one, you set:

• Network — which network the balancer works in.
• Pool members — the servers the traffic is distributed among.
• Listener — the protocol and port on which the balancer receives traffic (for example HTTP : 80).
• Members — the protocol and port to which it forwards traffic to the servers (for example HTTP : 80).

A typical scenario is a web farm: one public entry point, with several identical web servers behind it.

Security groups

A security group is a cloud firewall: it decides which inbound and outbound traffic is allowed to a server (ports, protocols, sources). This is a separate large topic — the article "Security groups: the cloud firewall" is devoted to it. In short: without an allow rule a server is unreachable from outside, and the typical ready-made groups are SSH, ICMP, default.

VPN (IPsec)

The VPNs section is a site-to-site VPN over the IPsec protocol: a secure tunnel between your cloud and another site (an office, another data center). It consists of several parts (sub-tabs):

• VPN Gateways — the VPN service on a router (the tunnel's entry point).
• IKE Policies and IPSec Policies — the encryption and key-negotiation parameters.
• VPN Endpoint Groups — which local and remote subnets are linked.
• IPSec Site Connections — the connection between the sites itself.

When you need it. An IPsec VPN is for a permanent "network-to-network" link (office ↔ cloud). If the task is to give employees remote access, it is often easier to set up your own VPN server (for example WireGuard) on an ordinary cloud server.

Typical scenarios

• A single server on the internet. A server in a public network + a security group with the needed ports (SSH/HTTP/HTTPS). A floating IP — if you need a fixed external address.
• A private network with outside access. Your own network + subnet, a router with an external gateway, floating IPs on the servers that must be visible from outside. The rest stay hidden.
• A fault-tolerant site. Several web servers + a load balancer in front of them; the security group opens 80/443 only to the balancer.
• Connecting an office to the cloud. An IPsec site-to-site VPN between the cloud router and the office gateway — internal resources become available without exposing them to the internet.

FAQ

What is the difference between a network and a subnet?

A network is a "switch" (layer L2) that servers connect to. A subnet is a range of IP addresses (layer L3) inside that network. A single network can have a subnet with a CIDR, gateway and DHCP; without a subnet servers have nowhere to get addresses.

How is a floating IP different from an ordinary public address?

A floating IP is not "nailed" to a specific server: it can be removed and assigned to another server in seconds. This is handy for migrations and fault tolerance — the external address stays the same.

Do I need a router if the server is already in a public network?

If the server is created directly in a public network (public2), it already has internet access — a separate router is not needed. A router is needed when you build your own private network and want to give it outside access.

Is it mandatory to create your own network?

No. For simple tasks a server can be created in a ready-made public network. Your own network is needed when you want to isolate servers, separate environments (web/DB) or build a more complex topology.

Is a load balancer the same as several floating IPs?

No. Floating IPs simply hand out different addresses to different servers. A load balancer provides one entry point and smartly distributes requests among servers, checking their availability.