A Virtual Private Network (VPN) is a computer network in which some of the links between nodes are carried by open connections or virtual circuits in some larger networks (such as the Internet), as opposed to running across a single private network. The Link Layer protocols of the virtual network are said to be tunneled through the transport network. One common application is to secure communications through the public Internet, but a VPN does not need to have explicit security features such as authentication or content encryption. For example, VPNs can also be used to separate the traffic of different user communities over an underlying network with strong security features, or to provide access to a network via customized or private routing mechanisms.

VPN service providers may offer best-effort performance, or may have a defined service level agreement (SLA) with their VPN customers. Generally, a VPN has a topology more complex than point-to-point.
The Internet Engineering Task Force (IETF) has categorized a variety of VPNs, some of which, such as Virtual LANs (VLAN) are the standardization responsibility of other organizations, such as the Institute of Electrical and Electronics Engineers (IEEE) Project 802, Workgroup 802.1 (architecture). Originally, Wide Area Network (WAN) links from a telecommunications service provider interconnected network nodes within a single enterprise. With the advent of LANs, enterprises could interconnect their nodes with links that they owned. While the original WANs used dedicated lines and layer 2 multiplexed services such as Frame Relay, IP-based layer 3 networks, such as the ARPANET, Internet, military IP networks (NIPRNET, SIPRNET, JWICS, etc.), became common interconnection media. VPNs began to be defined over IP networks. The military networks may themselves be implemented as VPNs on common transmission equipment, but with separate encryption and perhaps routers.

It became useful first to distinguish among different kinds of IP VPN based on the administrative relationships (rather than the technology) interconnecting the nodes. Once the relationships were defined, different technologies could be used, depending on requirements such as security and quality of service.
When an enterprise interconnects a set of nodes, all under its administrative control, through a LAN network, that is termed an intranet. When the interconnected nodes are under multiple administrative authorities but are hidden from the public Internet, the resulting set of nodes is called an extranet. A user organization can manage both intranets and extranets itself, or negotiate a service as a contracted (and usually customized) offering from an IP service provider. In the latter case, the user organization contracts for layer 3 services – much as it may contract for layer 1 services such as dedicated lines, or multiplexed layer 2 services such as frame relay.

IETF documents distinguish between provider-provisioned and customer-provisioned VPNs. Just as an interconnected set of providers can supply conventional WAN services, so a single service provider can supply provider-provisioned VPNs (PPVPNs), presenting a common point-of-contact to the user organization.

Tunneling protocols can be used in a point-to-point topology that would generally not be considered a VPN, because a VPN is expected to support arbitrary and changing sets of network nodes. Since most router implementations support software-defined tunnel interface, customer-provisioned VPNs often comprise simply a set of tunnels over which conventional routing protocols run. PPVPNs, however, need to support the coexistence of multiple VPNs, hidden from one another, but operated by the same service provider.

Depending on whether the PPVPN runs in layer 2 or layer 3, the building blocks described below may be L2 only, L3 only, or combinations of the two. Multiprotocol Label Switching (MPLS) functionality blurs the L2-L3 identity..

While RFC 4026 generalized these terms to cover L2 and L3 VPNs, they were introduced in RFC 2547.
In general, a CE is a device, physically at the customer premises, that provides access to the PPVPN service. Some implementations treat it purely as a demarcation point between provider and customer responsibility, while others allow customers to configure it.

A PE is a device or set of devices, at the edge of the provider network, which provides the provider's view of the customer site. PEs are aware of the VPNs that connect through them, and which maintain VPN state.
A P device operates inside the provider's core network, and does not directly interface to any customer endpoint. It might, for example, provide routing for many provider-operated tunnels that belong to different customers' PPVPNs. While the P device is a key part of implementing PPVPNs, it is not itself VPN-aware and does not maintain VPN state. Its principal role is allowing the service provider to scale its PPVPN offerings, as, for example, by acting as an aggregation point for multiple PEs. P-to-P connections, in such a role, often are high-capacity optical links between major locations of provider.
The customer determines the overall customer VPN service, which also can involve routing, bridging, or host network elements.

An unfortunate acronym confusion can occur between Virtual Private Line Service and Virtual Private LAN Service; the context should make it clear whether "VPLS" means the layer 1 virtual private line or the layer

2 virtual private LAN.
A Layer 2 technique that allows for the coexistence of multiple LAN broWNOast domains, interconnected via trunks using the IEEE 802.1Q trunking protocol. Other trunking protocols have been used but have become obsolete, including Inter-Switch Link (ISL), IEEE 802.10 (originally a security protocol but a subset was introduced for trunking), and ATM LAN Emulation (LANE).
Developed by IEEE, VLANs allow multiple tagged LANs to share common trunking. VLANs frequently comprise only customer-owned facilities. The former[clarification needed] is a layer 1 technology that supports emulation of both point-to-point and point-to-multipoint topologies. The method discussed here extends Layer 2 technologies such as 802.1d and 802.1q LAN trunking to run over transports such as Metro

As used in this context, a VPLS is a Layer 2 PPVPN, rather than a private line, emulating the full functionality of a traditional local area network (LAN). From a user standpoint, a VPLS makes it possible to interconnect several LAN segments over a packet-switched, or optical, provider core; a core transparent to the user, making the remote LAN segments behave as one single LAN.
In a VPLS, the provider network emulates a learning bridge, which optionally may include VLAN service.

PW is similar to VPWS, but it can provide different L2 protocols at both ends. Typically, its interface is a WAN protocol such as Asynchronous Transfer Mode or Frame Relay. In contrast, when aiming to provide the appearance of a LAN contiguous between two or more locations, the Virtual Private LAN service or IPLS would be appropriate.....
A subset of VPLS, the CE devices must have L3 capabilities; the IPLS presents packets rather than frames. It may support IPv4 or IPv6.

This section discusses the main architectures for PPVPNs, one where the PE disambiguates duplicate addresses in a single routing instance, and the other, virtual router, in which the PE contains a virtual router instance per VPN. The former approach, and its variants, have gained the most attention.
One of the challenges of PPVPNs involves different customers using the same address space, especially the IPv4 private address space. The provider must be able to disambiguate overlapping addresses in the multiple customers' PPVPNs.

In the method defined by RFC 2547, BGP extensions advertise routes in the IPv4 VPN address family, which are of the form of 12-byte strings, beginning with an 8-byte Route Distinguisher (RD) and ending with a 4-byte IPv4 address. RDs disambiguate otherwise duplicate addresses in the same PE.

PEs understand the topology of each VPN, which are interconnected with MPLS tunnels, either directly or via P routers. In MPLS terminology, the P routers are Label Switch Routers without awareness of VPNs.
The Virtual Router architecture, as opposed to BGP/MPLS techniques, requires no modification to existing routing protocols such as BGP. By the provisioning of logically independent routing domains, the customer operating a VPN is completely responsible for the address space. In the various MPLS tunnels, the different PPVPNs are disambiguated by their label, but do not need routing distinguishers.
Virtual router architectures do not need to disambiguate addresses, because rather than a PE router having awareness of all the PPVPNs, the PE contains multiple virtual router instances, which belong to one and only one VPN.

From the security standpoint, VPNs either trust the underlying delivery network, or must enforce security with mechanisms in the VPN itself. Unless the trusted delivery network runs only among physically secure sites, both trusted and secure models need an authentication mechanism for users to gain access to the VPN.

Some Internet service providers as of 2009[update] offer managed VPN service for business customers who want the security and convenience of a VPN but prefer not to undertake administering a VPN server themselves. Managed VPNs go beyond PPVPN scope, and are a contracted security solution that can reach into hosts. In addition to providing remote workers with secure access to their employer's internal network, other security and management services are sometimes included as part of the package. Examples include keeping anti-virus and anti-spyware programs updated on each client's computer.

A known trusted user, sometimes only when using trusted devices, can be provided with appropriate security privileges to access resources not available to general users. Servers may also need to authenticate themselves to join the VPN.

A wide variety of authentication mechanisms exist. VPNs may implement authentication in devices including firewalls, access gateways, and others. They may use passwords, biometrics, or cryptographic methods. Strong authentication involves combining cryptography with another authentication mechanism. The authentication mechanism may require explicit user action, or may be embedded in the VPN client or the workstation.

Trusted VPNs do not use cryptographic tunneling, and instead rely on the security of a single provider's network to protect the traffic. In a sense, they elaborate on traditional network- and system-administration work.

Multi-Protocol Label Switching (MPLS) is often used to overlay VPNs, often with quality-of-service control over a trusted delivery network.

Layer 2 Tunneling Protocol (L2TP) which is a standards-based replacement, and a compromise taking the good features from each, for two proprietary VPN protocols: Cisco's Layer 2 Forwarding (L2F) (obsolete as of 2009[update]) and Microsoft's Point-to-Point Tunneling Protocol (PPTP).

Secure VPNs use cryptographic tunneling protocols to provide the intended confidentiality (blocking intercept and thus packet sniffing), sender authentication (blocking identity spoofing), and message integrity (blocking message alteration) to achieve privacy.

Secure VPN protocols include the following:
IPsec (Internet Protocol Security) - A standards-based security protocol developed originally for IPv6, where support is mandatory, but also widely used with IPv4.
Transport Layer Security (SSL/TLS) is used either for tunneling an entire network's traffic (SSL VPN), as in the OpenVPN project, or for securing individual connection. SSL has been the foundation by a number of vendors to provide remote access VPN capabilities. A practical advantage of an SSL VPN is that it can be accessed from locations that restrict external access to SSL-based e-commerce websites without IPsec implementations. SSL-based VPNs may be vulnerable to Denial of Service attacks mounted against their TCP connections because latter are inherently unauthenticated.

DTLS, used by Cisco for a next generation VPN product called Cisco AnyConnect VPN. DTLS solves the issues found when tunneling TCP over TCP as is the case with SSL/TLS
Secure Socket Tunneling Protocol (SSTP) by Microsoft introduced in Windows Server 2008 and Windows Vista Service Pack 1. SSTP tunnels Point-to-Point Protocol (PPP) or L2TP traffic through an SSL 3.0 channel.

L2TPv3 (Layer 2 Tunneling Protocol version 3), a new[update] release.
MPVPN (Multi Path Virtual Private Network). Ragula Systems Development Company owns the registered trademark "MPVPN".
Cisco VPN, a proprietary VPN used by many Cisco hardware devices. Proprietary clients exist for all platforms; open-source clients also exist.
SSH VPN -- OpenSSH offers VPN tunneling to secure remote connections to a network (or inter-network links). This feature (option -w) should not be confused with port forwarding (option -L). OpenSSH server provides limited number of concurrent tunnels and the VPN feature itself does not support personal authentication.
Mobile VPNs apply standards-based authentication and encryption technologies to secure communications with mobile devices and to protect networks from unauthorized users. Designed for wireless environments, Mobile VPNs provide an access solution for mobile users who require secure access to information and applications over a variety of wired and wireless networks. Mobile VPNs allow users to roam seamlessly across IP-based networks and in and out of wireless-coverage areas without losing application sessions or dropping the secure VPN session. For instance, highway patrol officers require access to mission-critical applications as they travel between different subnets of a mobile network, much as a cellular radio has to hand off its link to repeaters at different cell towers.

The Host Identity Protocol (HIP), under study by the Internet Engineering Task Force, is designed to support mobility of hosts by separating the role of IP addresses for host identification from their locator functionality in an IP network. With HIP a mobile host maintains its logical connections established via the host identity identifier while associating with different IP addresses when roaming between access networks.