A Core Competence Product Evaluation

Switch Management

The Desktop Switch and Switch 1000 are SNMP-manageable and configurable. Using SunNet Manager and HP Open View for Windows, we performed network discovery and conducted polling, alarm and trap processing, and found the resulting level of manageability to be a substantial improvement over our previous hubbed environment.

In addition to SNMP MIB-II, both switches support the IETF bridge and repeater MIB modules and seven Remote Monitoring (RMON) MIB groups (alarms, events, statistics, history, host table, top ten hosts, and matrix). Of these, we believe RMON support is an increasingly important consideration for management as networks grow because of the reduced polling, centralization, and advanced alarm and monitoring RMON probes provide. RMON support is provided through probe functionality embedded in the SNMP agent on the SuperStack switches, which eliminates the need for a passive, stand-alone probe unit.

RMON works as follows. An RMON probe (hardware or software agent) passively collects information about the performance and status of a network segment(s), which can then be collected by an SNMP network management station directly from the probe as opposed to the individual end stations on the network segment(s). As an example, we include the results of a single SNMP get-request for the ifStatus table from our Switch 1000, executed using the Quick Dump facility from our Sun Net Manager NMS:

Mon Feb 24 10:54:26 1997 [ 207.86.152.190 ] : Quick Dump: snmp-mibII.ifStatus
KEY	ifIndex	ifDescr	ifSpeed	ifPhysAddress	ifAdminStatus	ifOperStatus
1	1	3Com LinkSwitch 1000	10000000	08:00:4E:12:27:CA	up	up
101	101	RMON:V2 Port 1 on Unit 1	10000000		up	down
102	102	RMON:V2 Port 2 on Unit 1	10000000		up	up
103	103	RMON:V2 Port 3 on Unit 1	10000000		up	up
104	104	RMON:V2 Port 4 on Unit 1	10000000		up	down
105	105	RMON:V2 Port 5 on Unit 1	10000000		up	up
106	106	RMON:V2 Port 6 on Unit 1	10000000		down	down
107	107	RMON:V2 Port 7 on Unit 1	10000000		up	up
108	108	RMON:V2 Port 8 on Unit 1	10000000		up	up
109	109	RMON:V2 Port 9 on Unit 1	10000000		up	up
110	110	RMON:V2 Port 10 on Unit 1	10000000		up	down
111	111	RMON:V2 Port 11 on Unit 1	10000000		up	up
112	112	RMON:V2 Port 12 on Unit 1	10000000		up	down
113	113	RMON:V2 Port 13 on Unit 1	10000000		up	down
114	114	RMON:V2 Port 14 on Unit 1	10000000		up	down
115	115	RMON:V2 Port 15 on Unit 1	10000000		up	up
116	116	RMON:V2 Port 16 on Unit 1	10000000		up	down
117	117	RMON:V2 Port 17 on Unit 1	10000000		up	down
118	118	RMON:V2 Port 18 on Unit 1	10000000		up	down
119	119	RMON:V2 Port 19 on Unit 1	10000000		up	down
120	120	RMON:V2 Port 20 on Unit 1	10000000		up	down
121	121	RMON:V2 Port 21 on Unit 1	10000000		up	down
122	122	RMON:V2 Port 22 on Unit 1	10000000		up	down
123	123	RMON:V2 Port 23 on Unit 1	10000000		up	up
124	124	RMON:V2 Port 24 on Unit 1	10000000		up	down
126	126	RMON:V3 Port 26 on Unit 1	100000000		up	up
1001	1001	RMON: VLAN 1	220000000	03:E9:00:00:00:00	up	up
1002	1002	RMON: VLAN 2	60000000	03:EA:00:C0:00:00	up	up
1003	1003	RMON: VLAN 3	60000000	03:EB:00:02:00:00	up	up
1004	1004	RMON: VLAN 4	0	03:EC:00:00:00:00	up	up
1005	1005	RMON: VLAN 5	0	03:ED:00:D0:00:0B	up	up
1006	1006	RMON: VLAN 6	0	03:EE:00:D0:00:00	up	up
1007	1007	RMON: VLAN 7	0	03:EF:00:D0:00:00	up	up
1008	1008	RMON: VLAN 8	0	03:F0:00:D0:00:00	up	up
1009	1009	RMON: VLAN 9	0	03:F1:00:02:00:00	up	up
1010	1010	RMON: VLAN 10	0	03:F2:01:02:00:00	up	up
1011	1011	RMON: VLAN 11	0	03:F3:00:00:00:00	up	up
1012	1012	RMON: VLAN 12	0	03:F4:00:D0:00:00	up	up
1013	1013	RMON: VLAN 13	0	03:F5:00:D0:00:00	up	up
1014	1014	RMON: VLAN 14	0	03:F6:00:00:00:00	up	up
1015	1015	RMON: VLAN 15	0	03:F7:00:02:00:00	up	up
1016	1016	RMON: VLAN 16	0	03:F8:00:28:00:00	up	up

RMON probes can be configured to generate alarms when certain threshold conditions are met (e.g., a dangerously high percentage of bandwidth in use on a LAN segment). By default, both switches enable four threshold-based alarms: bandwidth used, broadcast bandwidth used, percentage of packets forwarded, and errors per 10,000 packets.

All our configurations were entered through the menu-driven interfaces of these switches, which we found surprisingly easy to use. The screens are relatively clutter-free, and we were pleasantly surprised to find the menus highly consistent across the Desktop Switch and Switch 1000 (with the obvious exception of features that are unique to either). 3Com claims that configuration is greatly facilitated through the use of their Transcend Enterprise Manager, which we hope to evaluate at a future time.

Both the Desktop Switch and Switch 1000 provide a meaningful amount of management information through the Telnet and VT100 console management. Switch unit and per port traffic statistics are provided, including frame counts (received, forwarded, filtered), error counts (bad CRCs, runts, overruns, late events and jabbers). The frame size analysis, and per port running averages of bandwidth used, frames forwarded, broadcast frame bandwidth, and error frames are likely to be useful performance measurement tools. A remote poll or ping facility is available for IP and IPX protocols (we were disappointed but not entirely surprised that a remote poll for AppleTalk was not present). Both switches are remote software upgradable via TFTP.

next...

Want to know more about VLANs