Dear Yaron, Thank you for your comments! We have made an update based on your comments. Please suggest further text if required. The Security section is further expended and more details are provided, especially the new section on the Malicious PCC is added as below. 9.2. Malicious PCC The PCECC mechanism described in this document requires the PCE to keep labels (CCI) that it downloads and relies on the PCC responding (with either an acknowledgment or an error message) to requests for LSP instantiation. This is an additional attack surface by placing a requirement for the PCE to keep a CCI/label replica for each PCC. It is RECOMMENDED that PCE implementations provide a limit on resources (in this case the CCI) a single PCC can occupy. [RFC8231] provides a notification mechanism when such threshold is reached. Please also find the working copy and the diff. Working copy: https://raw.githubusercontent.com/dhruvdhody/ietf/master/draft-ietf-pce-pcep-extension-for-pce-controller-11.txt Diff: https://tools.ietf.org/tools/rfcdiff/rfcdiff.pyht?url1=draft-ietf-pce-pcep-extension-for-pce-controller-10&url2=https://raw.githubusercontent.com/dhruvdhody/ietf/master/draft-ietf-pce-pcep-extension-for-pce-controller-11.txt Best Regards, Shuping > -----Original Message----- > From: Yaron Sheffer via Datatracker [mailto:noreply@xxxxxxxx] > Sent: Sunday, February 7, 2021 3:35 AM > To: secdir@xxxxxxxx > Cc: draft-ietf-pce-pcep-extension-for-pce-controller.all@xxxxxxxx; > last-call@xxxxxxxx; pce@xxxxxxxx > Subject: Secdir last call review of > draft-ietf-pce-pcep-extension-for-pce-controller-10 > > Reviewer: Yaron Sheffer > Review result: Not Ready > > This document defines PCEP extensions for the RFC 8283 architecture, where > the PCE acts as a central controller in an SDN. The document is focused on > specific use cases, referred to as "basic PCECC mode". > > Let me state up front that I am not familiar with the PCE architecture other > than what I read up in order to review this document. Having said that, I > suspect that there would be significant value in a security analysis of the > architecture defined here. Having each connection "authenticated and > encrypted" > is table stakes nowadays, but is it really enough for very large SDN > deployments that require this level of protocol sophistication? > > Details > > 9.1: "authenticated and encrypted" TLS sessions are typically only > authenticated by the server. Please point out explicitly that mutual > authentication is required. Also, is there no authorization? I would assume a > peer PCE Controller is allowed to do different things than a PCC. Are all PCCs > allowed to issue the same commands/queries, targeted at the same > resources? > > - RFC 8283 which defines the architecture that is implemented by this draft > says: > > [The] security implications of SDN have not been fully discussed or > described. > Therefore, protocol and applicability work-around solutions for this > architecture must take proper account of these concerns. > > It is expected that each new document that is produced for a specific use > case will also include considerations of the security impacts of the use of a > PCE-based central controller on the network type and services being > managed. > > I don't think that the current document addresses this challenge. > > In general, this looks like a very monolithic architecture, where everybody > trusts everybody else once they've been authenticated. Although Sec. 9.1 > discusses the case of a malicious PCE (which would be rather catastrophic), I > would encourage the authors to consider whether a malicious PCC can also > disrupt the PCE's operations and cause "major impact to the network". > -- last-call mailing list last-call@xxxxxxxx https://www.ietf.org/mailman/listinfo/last-call
