User Tools

Site Tools


wiki:rrcp_security

RRCP security model

Abstract

RRCP is a protocol for managing switches. Since such switch may be exposed to customers, peer partners or otherwise not-trusted envirounment, security must be considered in regard of controlling who can manage the switch, and who - not.

Description

OpenRRCP can enable or disable RRCP protocol on per-port basis.

RRCP is L2 protocol and, thus cannot propagate through L3 envirounment such as router or L3 switches. VLAN's are also efficient to isolate RRCP-enabled managing links from untrusted customer-data links.

Additionaly, AuthKey parameter can be used for further security escalation. However, unfortunately, this parameter is only 16-bit long and thus, all of it 65536 possible values can be brute-forced in several minutes.

To make AuthKey protection strong enough, administrator need to change switch MAC address and disable REP protocol. In this mode 48-bit(MAC)+16-bit(AuthKey) = 64 Bit brute-force attack will be required to compromise switch. Since 2^64 iterations of minimum 1 millisecond each requires more than 500 millions years to complete, this protection is considered reasonable.

Model

wiki/rrcp_security.txt · Last modified: 2015/11/08 13:07 (external edit)