* Define XML Schemas for SNMP message format, including Trap, Get, GetNext, and MetaDataQuery.
* Define SNMP specific WSDL extensions.
* SNMP BC should support device learning Performance Management(PM) and MOF
* Support "in-only" MEP for incoming traps, and "request-reply" MEP for MetaDataQuery and GET requests.
* Engine should process high volumn of incoming Traps through use of buffering, batching.
* Hookup to CAM specific monitoring MBeans.
* Extend runtime configuration MBeans for SNMP Engine parameters, such as batch size, wait-time, etc.
* Throttling of incoming Traps to avoid overloading Processor Engines.
* MetaData cache expiration
* Should actively monitor variables up to 100,000 network elements using SNMP GET
* Receive SNMP TRAP, INFORM, should be able to receive these traps at very high burst rates, buffer them and deliver them to Processing SEs.
* Dynamic configuration: addition / changes in the configuration of what to monitor without redeploying
* Dynamic addition of Processing Engines
Get, GetNext, or Set, TRAP
Get, Get Next, Set Request, Get Response, Notifications (TRAP), GETBULK, INFORM
support security in SNMPv2.
* support for SNMP operations such as Inform, Set, GetBulk, Walk, etc.
* Security and Remote Configuration
* Message integrity to ensure that a packet has not been tampered with in transit.
* Authentication to verify that the message is from a valid source.
* Encryption of packets to prevent snooping by an unauthorized source.
* Monitoring device uptimes (sysUpTimeInstance)
* Inventory of OS versions (sysDescr)
* Collect interface information (ifName, ifDescr, ifSpeed, ifType, ifPhysAddr)
* Measuring network interface throughput (ifInOctets, ifOutOctets)
* Querying a remote ARP cache (ipNetToMedia)
* Need to account for Index shifting behaviour
(Need to see how we can handle Index Shifting)
* Demonstrate the ability to receive multiple SNMP traps and transform these to alarms that are written to the alarm output file.
* Can we have a Wizard to create device views? (Using Discovery protocol)
* Provision to log traps
* Do we need to support syslog?
* Generic way of handling SNMP extensions
* Any additional data added to the trap must not be discarded.
* Querying specific agents to ensure they are still functioning