Hi Stephen,

Thanks for your inputs  vis-a-vis  4301/2/3. I concur that IPSec is not just 
about encryption but don't quite buy into what somebody spelled out during the 
meeting as 'IPSec is an access control mechanism that also provides other 
security services'; IMO, strict access control is more a firewall 
functionality. RFC 4301 spells out the access control rationale as
"IPsec includes a specification for minimal firewall functionality, since that 
is an essential aspect of access control at the IP layer… The IPsec firewall 
function makes use of the cryptographically-enforced authentication and 
integrity provided for all IPsec traffic to offer better access control than 
could be obtained through use of a firewall (one not privy to IPsec internal 
parameters) plus separate cryptographic protection."

I know that we might have ruffled a few feathers wrt making the SPD relatively 
trivial but let's get down to some details as far as the comparison goes. The 
first ADVPN proposal does talk about the shortcut suggester possibly including 
traffic selectors in the shortcut exchange. While this seems to give the notion 
of the ability to provide SA granularity, the source of such information is a 
third party (even though both peers have an active connection with this third 
party) and doesn't quite stand up to the very reason of including access 
control in IPSec (The IPsec firewall function makes use of the 
cryptographically-enforced authentication and integrity provided for all IPsec 
traffic to offer better access control than could be obtained through use of a 
firewall (one not privy to IPsec internal parameters) plus separate 
cryptographic protection.) - this is a case of the information not even privy 
to the same device, leave alone the same layer in the same device.

Let's say, the shortcut suggester passes on an information in the shortcut 
exchange to the two peers about the selectors (X->Y) to be used for the 
shortcut while selectors (X'->Y') be still allowed through the spoke-hub-spoke 
path. It's entirely upto the peers (shortcut partners) to choose what to do 
with that; well, the draft prescribes that the shortcut partners should use 
this but when we are talking about multi-vendor, multiple administrative 
domains, how do we trust that the shortcut partners do that. We must realise 
that the role of the suggester/hub is more advisory in nature; it can only 
advise not enforce. The problem is a system wide design problem not one to be 
sorted out just between two nodes (and IKE is intrinsically designed to operate 
in a peer-to-peer model), so e.g. even if policies prescribe (as obtained from 
the shortcut exchange) that a certain traffic be routed via the hub, once the 
two peers(shortcut partners) have all the authentication information for each 
other, they might give a damn and say we don't care, lets bypass the policy and 
use the shortcut for everything leaving the hub in dark. Unlike a traditional 
IPSec case, where each peer could enforce specific access control policies 
(e.g. if the peer you negotiated an SA with sends a traffic using an SA which 
was not supposed to be used for that traffic, you could simply drop it post 
decryption), the hub can't enforce such policies – it can dictate what doesn't 
pass through it but not what MUST pass through it, which is what the common 
requirement is.

What all this effectively means is if really need to enforce access-control 
policies, it has to be agreed between the spokes/shortcut partners, not 
prescribed by the hub. IPSec typically deals with a case where the SPD is in 
place apriori based on which SAs are negotiated but here we end up with a case 
where the SPDs need to be negotiated and agreed upon as well !!! In that sense, 
I would say DMVPN still provides better access control – it already determines 
what portion of the network should use the shortcut depends on the 
address/subnet sent in NHRP resolution reply. This is used to install shortcut 
routes and/or policy based routes and the same can be used to also install an 
ACL if we suspect the peer to play foul. It's another matter that this is not 
as part of IPSec access control directly. This is on top of the 
route/forwarding policy based tunnel selection.

- Manish

From: Stephen Kent <[email protected]<mailto:[email protected]>>
Date: Tuesday, 5 November 2013 3:27 AM
To: Mike Sullenberger <[email protected]<mailto:[email protected]>>, Michael 
Richardson <[email protected]<mailto:[email protected]>>
Cc: "Stephen Lynn (stlynn)" <[email protected]<mailto:[email protected]>>, 
"[email protected]<mailto:[email protected]>"
 
<[email protected]<mailto:[email protected]>>,
 Mark Comeadow <[email protected]<mailto:[email protected]>>, "Michael 
Guilford (mguilfor)" <[email protected]<mailto:[email protected]>>, IPsecme 
WG <[email protected]<mailto:[email protected]>>
Subject: Re: [IPsec] AD VPN: discussion kick off
Resent-From: 
<[email protected]<mailto:[email protected]>>
Resent-To: Frederic Detienne <[email protected]<mailto:[email protected]>>, Manish 
Kumar <[email protected]<mailto:[email protected]>>, Mike Sullenberger 
<[email protected]<mailto:[email protected]>>
Resent-Date: Tuesday, 5 November 2013 3:27 AM

Mike,

A couple of your comments caught my attention, as an author of 4301, 02, and 
03. I admit to not having read the DMVPN proposal, so my comments are based 
only on your message, which argues why DMVPN is the preferred solution.
IPsec encryption layer.  In this layer ISAKMP/IKEv2/IPsec is the correct 
standard protocol to use.  This is what IPsec does really well, encrypt 
traffic. The layers above greatly simplifies IPsec's job by presenting to it 
the tunnel to encrypt instead of all of the individual protocols/subnets/flows 
that are within the tunnel.  The IPsec selectors are now for the tunnel, which 
makes path redundancy and load-balancing  doable. IPsec doesn't deal well with 
the same set of selectors to encrypt traffic to more than one peer.  With DMVPN 
this is handled at the routing/forwarding and GRE tunnel layers.
IPsec is not just about encryption, although the DMVPN proposal may relegate it 
tothat. IPsec provides access control, and, typically, authentication.  Does 
DMVPN preserve the access control features of IPsec, or are users now relying 
on a hub to do this, or what?
 ...  With 10s of thousands of nodes and perhaps 100s of thousands of 
network/subnets reachable via the VPN the number of IPsec selectors across the 
VPN would get completely out of hand, especially if each different pair of 
subnets (selector) requires a separate encryption, between the same two nodes.
More properly, a separate SA, and only if the folks who manage policies at each 
end of the SA decide to provide fine-grained access control for the traffic 
flows. It was not clear to me that the problem statement for this work 
envisioned VPNs of the scale you mention. Also, the comments above are a bit 
confusing. Both end points and security gateways are "nodes" wrt IPsec, in the 
general sense. I can create a selector that secures traffic from my node (and 
point of subnet) to all hosts on a subnet, irrespective of how many are present 
there.
This doesn’t even count the fact that in order to run IPv4 and IPv6 between the 
same two nodes you have to use at least double the number of selectors.
At least? Under what circumstances would the number grow by more than a factor 
of two?
Routing protocols are already designed to scale to 100s of thousands and even 
millions of routes. So with DMVPN the forwarding and GRE tunneling of both IPv4 
and IPv6 is handled within a single GRE tunnel and IPsec selector.
So, the proposal simplifies use of IPsec by limiting the granularity at which 
SAs may be created?Does it also cause each SA to terminate at a hub, so that 
the security services are no longer e-t-e?  In the context of the perpass 
discussions, this seems like a questionable design decision.

Steve
_______________________________________________
IPsec mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/ipsec

Reply via email to