| IPsecME Working Group | V. Manral |
| Internet-Draft | HP |
| Intended status: Informational | S. Hanna |
| Expires: January 16, 2014 | Juniper |
| July 15, 2013 |
Auto Discovery VPN Problem Statement and Requirements
draft-ietf-ipsecme-ad-vpn-problem-09
This document describes the problem of enabling a large number of systems to communicate directly using IPsec to protect the traffic between them. It then expands on the requirements, for such a solution.
Manual configuration of all possible tunnels is too cumbersome in many such cases. In other cases the IP address of endpoints change or the endpoints may be behind NAT gateways, making static configuration impossible. The Auto Discovery VPN solution will address these requirements.
This Internet-Draft is submitted in full conformance with the provisions of BCP 78 and BCP 79.
Internet-Drafts are working documents of the Internet Engineering Task Force (IETF). Note that other groups may also distribute working documents as Internet-Drafts. The list of current Internet-Drafts is at http://datatracker.ietf.org/drafts/current/.
Internet-Drafts are draft documents valid for a maximum of six months and may be updated, replaced, or obsoleted by other documents at any time. It is inappropriate to use Internet-Drafts as reference material or to cite them other than as "work in progress."
This Internet-Draft will expire on January 16, 2014.
Copyright (c) 2013 IETF Trust and the persons identified as the document authors. All rights reserved.
This document is subject to BCP 78 and the IETF Trust's Legal Provisions Relating to IETF Documents (http://trustee.ietf.org/license-info) in effect on the date of publication of this document. Please review these documents carefully, as they describe your rights and restrictions with respect to this document. Code Components extracted from this document must include Simplified BSD License text as described in Section 4.e of the Trust Legal Provisions and are provided without warranty as described in the Simplified BSD License.
IPsec [RFC4301] is used in several different cases, including tunnel-mode site-to-site VPNs and Remote Access VPNs. Both tunneling modes for IPsec gateways and host-to-host transport mode are supported in this document.
The subject of this document is the problem presented by large scale deployments of IPsec and the requirements on a solution to address the problem. These may be a large collection of VPN gateways connecting various sites, a large number of remote endpoints connecting to a number of gateways or to each other, or a mix of the two. The gateways and endpoints may belong to a single administrative domain or several domains with a trust relationship.
Section 4.4 of RFC 4301 describes the major IPsec databases needed for IPsec processing. It requires an extensive configuration for each tunnel, so manually configuring a system of many gateways and endpoints becomes infeasible and inflexible.
The difficulty is that a lot of configuration mentioned in RFC 4301 is required to set up a Security Association. IKE implementations need to know the identity and credentials of all possible peer systems, as well as the addresses of hosts and/or networks behind them. A simplified mechanism for dynamically establishing point-to-point tunnels is needed. Section 2 contains several use cases that motivate this effort.
ADVPN - Auto Discovery Virtual Private Network (ADVPN) is VPN solution that enables a large number of systems to communicate directly, with minimal configuration and operator intervention using IPsec to protect communication between them.
Endpoint - A device that implements IPsec for its own traffic but does not act as a gateway.
Gateway - A network device that implements IPsec to protect traffic flowing through the device.
Point-to-Point - Communication between two parties without active participation (e.g. encryption or decryption) by any other parties.
Hub - The central point in a star topology/ dynamic full mesh topology, or one of the central points in the full mesh style VPN, i.e. gateway where multiple other hubs or spokes connect to. The hubs usually forward traffic coming from encrypted links to other encrypted links, i.e. there are no devices connected to it in clear.
Spoke - The endpoint in a star topology/ dynamic full mesh topology, or gateway which forwards traffic from multiple cleartext devices to other hubs or spokes, and some of those other devices are connected to it in clear (i.e. it encrypts data coming from cleartext devices and forwards it to the ADVPN).
ADVPN Peer - any member of an ADVPN including gateways, endpoints, hub or spoke.
Star topology - This is the topology where there is direct connectivity only between the hub and spoke and communication between the 2 spokes happens through the hub.
Allied and Federated Environments - Environments where we have multiple different organizations that have close association and need to connect to each other.
Full Mesh topology - This is the topology where there is a direct connectivity between every Spoke to every other Spoke directly, without the traffic between the spokes having to be redirected through an intermediate hub device.
Dynamic Full Mesh topology - This is the topology where direct connections exist in a hub and spoke manner, but dynamic connections are created/ removed between the spokes on a need basis.
Security Association (SA) - Defined in [RFC4301].
The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in [RFC2119].
This section presents the key use cases for large-scale point-to-point VPN.
In all of these use cases, the participants (endpoints and gateways) may be from a single organization (administrative domain) or from multiple organizations with an established trust relationship. When multiple organizations are involved, products from multiple vendors are employed so open standards are needed to provide interoperability. Establishing communications between participants with no established trust relationship is out of scope for this effort.
Two endpoints wish to communicate securely via a point-to-point Security Association (SA).
The need for secure endpoint-to-endpoint communications is often driven by a need to employ high-bandwidth, low -latency local connectivity instead of using slow, expensive links to remote gateways. For example, two users in close proximity may wish to place a direct, secure video or voice call without needing to send the call through remote gateways, which would add latency to the call, consume precious remote bandwidth, and increase overall costs. Such a use case also enables connectivity when both users are behind NAT gateways. Such a use case ought to allow for seamless connectivity even as endpoints roam, even if they are moving out from behind a NAT gateway, from behind one NAT gateway to behind another, or from a standalone position to behind a NAT gateway.
In a star topology, when two endpoints communicate they need a mechanism for authentication, such that they do not expose themselves to impersonation by the other spoke endpoint.
A typical Enterprise traffic model follows a star topology, with the gateways connecting to each other using IPsec tunnels.
However for voice and other rich media traffic that requires a lot of bandwidth or is performance sensitive, the traffic tromboning (taking a suboptimal path) to the hub can create traffic bottlenecks on the hub and can lead to an increase in cost. A fully meshed solution would make best use of the available network capacity and performance but the deployment of a fully meshed solution involves considerable configuration, especially when a large number of nodes are involved. It is for this purpose spoke-to-spoke tunnels are dynamically created and torn-down. For the reasons of cost and manual error reduction, it is desired that there be minimal configuration on each gateway.
The solution ought to work in cases where the endpoints are in different administrative domains, albeit, ones that have an existing trust relationship (for example two organisations who are collaborating on a project, they may wish to join their networks, whilst retaining independent control over configuration). It is highly desirable that the solution works for the star, full mesh as well as dynamic full mesh topology.
The solution ought to also address the case where gateways use dynamic IP addresses.
Additionally, the routing implications of gateway-to-gateway communication need to be addressed. In the simple case, selectors provide sufficient information for a gateway to forward traffic appropriately. In other cases, additional tunneling (e.g., Generic Routing Encapsulation - GRE) and routing (e.g., Open Shortest Path First - OSPF) protocols are run over IPsec tunnels, and the configuration impact on those protocols needs to be considered. There is also the case when Layer-3 Virtual Private Networks (L3VPNs) operate over IPsec Tunnels.
When two gateways communicate, they need to use a mechanism for authentication, such that they do not expose themselves to the risk of impersonation by the other entities.
A mobile endpoint ought to be able to use the most efficient gateway as it roams in the internet.
A mobile user roaming on the Internet may connect to a gateway, which because of roaming is no longer the most efficient gateway to use (reasons could be cost/ efficiency/ latency or some other factor). The mobile user ought to be able to discover and then connect to the current most efficient gateway in a seamless way without having to bring down the connection.
Several solutions exist for the problems described above. However, none of these solutions is adequate, as described here.
One simple solution is to configure all gateways and endpoints in advance with all the information needed to determine which gateway or endpoint is optimal and to establish an SA with that gateway or endpoint. However, this solution does not scale in a large network with hundreds of thousands of gateways and endpoints, especially when multiple administrative domains are involved and things are rapidly changing (e.g. mobile endpoints). Such a solution is also limited by the smallest endpoint/ gateway, as the same exhaustive configuration is to be applied on all endpoints/ gateways. A more dynamic, secure and scalable system for establishing SAs between gateways is needed.
The most common way to address a part of this this problem today is to use what has been termed a "star topology". In this case one or a few gateways are defined as "hub gateways", while the rest of the systems (whether endpoints or gateways) are defined as "spokes". The spokes never connect to other spokes. They only open tunnels with the hub gateways. Also for a large number of gateways in one administrative domain, one gateway may be defined as the hub, and the rest of the gateways and remote access clients connect only to that gateway.
This solution however is complicated by the case when the spokes use dynamic IP addresses and DNS with dynamic updates needs to be used. It is also desired that there is minimal to no configuration on the hub as the number of spokes increases and new spokes are added and deleted randomly.
Another problem with the star topology is that it creates a high load on the hub gateways as well as on the connection between the spokes and the hub. This load is both in processing power and in network bandwidth. A single packet in the hub-and-spoke scenario can be encrypted and decrypted multiple times. It would be much preferable if these gateways and clients could initiate tunnels between them, bypassing the hub gateways. Additionally, the path bandwidth to these hub gateways may be lower than that of the path between the spokes. For example, two remote access users may be in the same building with high-speed wifi (for example, at an IETF meeting). Channeling their conversation through the hub gateways of their respective employers seems extremely wasteful, as well as having lower bandwidth.
The challenge is to build large scale, IPsec-protected networks that can dynamically change with minimal administrative overhead.
Several vendors offer proprietary solutions to these problems. However, these solutions offer no interoperability between equipment from one vendor and another. This means that they are generally restricted to use within one organization, and it is harder to move off such solutions as the features are not standardized. Besides multiple organizations cannot be expected to all choose the same equipment vendor.
This section defines the requirements, on which the solution will be based.
1. For any network topology (star, full mesh and dynamic full mesh), when a new gateway or endpoint is added, removed, or changed, configuration changes are minimized as follows. Adding or removing a spoke in the topology MUST NOT require configuration changes to the hubs other than where the spoke was connected to and SHOULD NOT require configuration changes to the hub the spoke was connected to. The changes also MUST NOT require configuration changes in other spokes.
Specifically, when evaluating potential proposals, we will compare them by looking at how many endpoints or gateways must be reconfigured when a new gateway or endpoint is added, removed, or changed and how substantial this reconfiguration is, besides the amount of static configuration required.
This requirement is driven by use cases 2.1 and 2.2 and by the scaling limitations pointed out in section 3.1.
2. ADVPN peers MUST allow IPsec Tunnels to be setup with other members of the ADVPN without any configuration changes, even when peer addresses get updated every time the device comes up. This implies that SPD entries or other configuration based on peer IP address will need to be automatically updated, avoided, or handled in some manner to avoid a need to manually update policy whenever an address changes.
3. In many cases additional tunneling protocols (e.g. GRE) or Routing protocols (e.g. OSPF) are run over the IPsec tunnels. Gateways MUST allow for the operation of tunneling and Routing protocols operating over spoke-to-spoke IPsec Tunnels with minimal or no, configuration impact. The ADVPN solution SHOULD NOT increase the amount of information required to configure protocols running over IPsec tunnels.
4. In the full mesh and dynamic full mesh topology, Spokes MUST allow for direct communication with other spoke gateways and endpoints. In the star topology mode, direct communication between spokes MUST be disallowed.
This requirement is driven by use cases 2.1 and 2.2 and by the limitations of a star topology pointed out in section 3.2.
5. Any of the ADVPN Peers MUST NOT have a way to get the long term authentication credentials for any other ADVPN Peers. The compromise of an Endpoint MUST NOT affect the security of communications between other ADVPN Peers. The compromise of a Gateway SHOULD NOT affect the security of the communications between ADVPN Peers not associated with that Gateway.
This requirement is driven by use case 2.1. ADVPN Peers (especially Spokes) become compromised fairly often. The compromise of one ADVPN Peer SHOULD NOT affect the security of other unrelated ADVPN Peers.
6. Gateways SHOULD allow for seamless handoff of sessions in case endpoints are roaming, even if they cross policy boundaries. This would mean the data traffic is minimally affected even as the handoff happens. External factors like firewall, NAT boxes that will be part of the overall solution when DVPN is deployed will not be considered part of this solution.
Such endpoint roaming may affect not only the endpoint-to-endpoint SA but also the relationship between the endpoints and gateways (such as when an endpoint roams to a new network that is handled by a different gateway).
This requirement is driven by use case 2.1. Today's endpoints are mobile and transition often between different networks (from 4G to WiFi and among various WiFi networks).
7. Gateways SHOULD allow for easy handoff of a session to another gateway, to optimize latency, bandwidth, load balancing, availability, or other factors, based on policy.
This ability to migrate traffic from one gateway to another applies regardless of whether the gateways in question are hubs or spokes. It even applies in the case where a gateway (hub or spoke) moves in the network, as may happen with a vehicle-based network.
This requirement is driven by use case 2.3.
8. Gateways and endpoints MUST have the capability to participate in an ADVPN even when they are located behind NAT boxes. However, in some cases they may be deployed in such a way that they will not be fully reachable behind a NAT box. It is especially difficult to handle cases where the Hub is behind a NAT box. Where the two endpoints are both behind separate NATs, communication between these spokes SHOULD be supported using workarounds such as port forwarding by the NAT or detecting when two spokes are behind uncooperative NATs and using a hub in that case.
This requirement is driven by use cases 2.1 and 2.2. Endpoints are often behind NATs and gateways sometimes are. IPsec SHOULD continue to work seamlessly regardless, using ADVPN techniques whenever possible and providing graceful fallback to hub and spoke techniques as needed.
9. Changes such as establishing a new IPsec SA SHOULD be reportable and manageable. However, creating a MIB or other management technique is not within scope for this effort.
This requirement is driven by manageability concerns for all the use cases, especially use case 2.2. As IPsec networks become more dynamic, management tools become more essential.
10. To support allied and federated environments, endpoints and gateways from different organizations SHOULD be able to connect to each other.
This requirement is driven by demand for all the use cases in federated and allied environments.
11. The administrator of the ADVPN SHOULD allow for the configuration of a Star, Full mesh or a partial full mesh topology, based on which tunnels are allowed to be setup.
This requirement is driven by demand for all the use cases in federated and allied environments.
12. The ADVPN solution SHOULD be able to scale for multicast traffic.
This requirement is driven by the use case 2.2, where the amount of rich media multicast traffic is increasing.
13. The ADVPN solution SHOULD allow for easy monitoring, logging and reporting of the dynamic changes, to help for trouble shooting such environments.
This requirement is driven by demand for all the use cases in federated and allied environments.
14. There is also the case when L3VPNs operate over IPsec Tunnels, for example Provider Edge (PE) based VPN's. An ADVPN MUST support L3VPN as an application protected by the IPsec Tunnels.
This requirement is driven by demand for all the use cases in federated and allied environments.
15. There ADVPN solution SHOULD allow the enforcement of per peer QoS in both the Star as well as the Full Mesh topology.
This requirement is driven by demand for all the use cases in federated and allied environments.
16. There ADVPN solution SHOULD take care of not letting the Hub to be a single point of failure.
This requirement is driven by demand for all the use cases in federated and allied environments.
This is a Problem statement and requirement document for the ADVPN solution, and in itself does not introduce any new security concerns. The solution to the problems presented in this draft may involve dynamic updates to databases defined by RFC 4301, such as the Security Policy Database (SPD) or the Peer Authorization Database (PAD).
RFC 4301 is silent about the way these databases are populated, and it is implied that these databases are static and pre-configured by a human. Allowing dynamic updates to these databases must be thought out carefully, because it allows the protocol to alter the security policy that the IPsec endpoints implement.
One obvious attack to watch out for is stealing traffic to a particular site. The IP address for www.example.com is 192.0.2.10. If we add an entry to an IPsec endpoint's SPD that says that traffic to 192.0.2.10 is protected through peer Gw-Mallory, then this allows Gw-Mallory to either pretend to be www.example.com or to proxy and read all traffic to that site. Updates to this database requires a clear trust model.
Hubs can be a single point of failure which can cause loss of connectivity of the entire system, that can be a big security issue. Any ADVPN solution design should take care of these concerns.
No actions are required from IANA for this informational document.
Many people have contributed to the development of this problem statement and many more will probably do so before we are done with it. While we cannot thank all contributors, some have played an especially prominent role. Yoav Nir, Yaron Sheffer, Jorge Coronel Mendoza, Chris Ulliott, and John Veizades wrote the document upon which this draft was based. Geoffrey Huang, Toby Mao, Suresh Melam, Praveen Sathyanarayan, Andreas Steffen, Brian Weis, and Lou Berger provided essential input.
| [RFC4301] | Kent, S. and K. Seo, "Security Architecture for the Internet Protocol", RFC 4301, December 2005. |
| [RFC2119] | Bradner, S., "Key words for use in RFCs to Indicate Requirement Levels", BCP 14, RFC 2119, March 1997. |