MIB-II Sample Clauses

MIB-II. ‌ The IETF defined MIB-II [RFC1213] as a MIB that all IP compliant devices should implement. The document has the status of ‘full standard’, which is the highest level in the IETF process of defining standards. MIB-II has experienced wide deployment in IP capable equipment. Furthermore, all the groups of management objects in MIB-II are mandatory; there are no optional objects defined. It is the combination of these two facts that make MIB-II a good candidate for retrieving information about SLA conformance of IP systems. Since the publication of this standard the management information it contains has been updated once more, and the single document was split into a set of separate documents [refs, to, them]. Since this happened only fairly recently, at this date the new set of doc- uments is not experiencing the wide deployment in a similar way as MIB-II. MIB-II organizes its management information into ten groups of objects, these are the system, inter- faces, at (address translation), icmp, tcp, udp, egp (exterior gateway protocol), dot3 (transmission) and snmp (simple network management protocol) groups. The objects in the various groups of MIB-II can be used to verify some of the SLA parameters defined in section 5. How this can be done and what particular MIB objects relate to what particular SLA is discussed next. The IP packet transfer delay nor the variation in packet delay can be verified using MIB-II objects. The detailed timing requirements that would be needed for this type of verification are not present anywhere in the MIB. The packet error ratio and the packet loss ratio can to some extent be verified by examining the rel- evant MIB objects in the IP group and the interfaces group. An estimated value for the error ratio at a given interface can be found by dividing the ifInErrors value for that interface by the sum of ifInU- castPkts and ifInNUcastPkts. Although some information can of course be attributed to this value, it still is just a poor estimate, for the following two reasons. First, the counter values are determined only at a single point in the network. The definition of the error ratio SLA performance parameter is in fact a end-to-end parameter, not a parameter at a single point. Problems elsewhere in the network cannot be detected at the measurement point, while at the same time those problems will affect the real value of the end-to-end performance parameter. Second, the counters count packets, not octets. Differences in size b...