Bandwidth lower speed then the port channel

[tudor347]

Hello, can anyone please tell me what could be the reason: 

We have a VM ware that is connected to a stack of two switches C9300L-48T-4X. 

On the Vmware and Switches there is created a port channel for link aggregation when we try to move data from the vmware trough the switch the speed is half of the port channel bandwidth. 

interface Port-channel7
description --- PO ESX 10 6G ---
switchport trunk allowed vlan 2,3,20,22
switchport mode trunk
logging event trunk-status
spanning-tree portfast trunk

Port-channel7 is up, line protocol is up (connected)
Hardware is EtherChannel, address is dc05.3926.9b0c (bia dc05.3926.9b0c)
Description: --- PO ESX 10 6G ---
MTU 1500 bytes, BW 4000000 Kbit/sec, DLY 10 usec,
reliability 255/255, txload 1/255, rxload 1/255
Encapsulation ARPA, loopback not set
Keepalive set (10 sec)
Full-duplex, 1000Mb/s, link type is auto, media type is N/A
input flow-control is on, output flow-control is unsupported
Members in this channel: Gi1/0/12 Gi1/0/13 Gi2/0/12 Gi2/0/13
ARP type: ARPA, ARP Timeout 04:00:00
Last input never, output 00:00:00, output hang never
Last clearing of "show interface" counters never
Input queue: 0/2000/0/0 (size/max/drops/flushes); Total output drops: 119502
Queueing strategy: fifo
Output queue: 0/40 (size/max)
5 minute input rate 4171000 bits/sec, 472 packets/sec
5 minute output rate 1098000 bits/sec, 374 packets/sec
18667217617 packets input, 27314412980145 bytes, 0 no buffer
Received 2959371 broadcasts (2194864 multicasts)
0 runts, 0 giants, 0 throttles
65 input errors, 10 CRC, 0 frame, 0 overrun, 0 ignored
0 watchdog, 2194864 multicast, 0 pause input
0 input packets with dribble condition detected
3266899178 packets output, 584277515589 bytes, 0 underruns
Output 53866653 broadcasts (38476039 multicasts)
0 output errors, 0 collisions, 1 interface resets
0 unknown protocol drops
0 babbles, 0 late collision, 0 deferred
0 lost carrier, 0 no carrier, 0 pause output
0 output buffer failures, 0 output buffers swapped out

Does anyone know if the bandwidth is split in the port channel between the VLANS? 

When we tried to transfer the data we tried to transfer only from a machine that is in VLAN2 in the vmware. 

Thank you

[Defututus]

The load balancing method depends on configured load balancing hash which is configured per box in your case. I'm not sure but I think that the default one is src-mac in case of C9300. It means that the speed will NOT exceed 1Gbit for single VM if is represented by a single MAC address. Here is a very good reading about Etherchannel load balancing: https://study-ccnp.com/etherchannel-load-balancing-explanation-configuration/. I Hope it helped

[MHM Cisco World]

Speed of PO =Speed of any port member 

BW of PO = to sum of BW of all ports member

Speed remains same freind in PO what change is BW

[M02@rt37]

hello @tudor347,

The bandwidth of a port channel is indeed aggregated and shared among the member links. Each member link contributes to the overall bandwidth of the port channel, and traffic is load balanced across the member links.

Based on the output you provided, the port channel "Port-channel7" has a bandwidth of 4 Gbps, which is the sum of the bandwidths of the four member links (Gi1/0/12, Gi1/0/13, Gi2/0/12, Gi2/0/13).

When transferring data through the port channel, the traffic should be distributed across the member links, allowing for higher overall throughput. However, keep in mind that a single flow or session will still be limited to the bandwidth of a single member link. This means that a single flow cannot exceed the bandwidth of one of the member links (in this case, 1 Gbps).

[LockStock2SmokingSwitches]

Based on the provided information, it seems that the port channel (Port-channel7) on the C9300L-48T-4X switches is operating at a speed of 1 Gbps (1000 Mb/s) and is configured as a trunk, allowing VLANs 2, 3, 20, and 22. The port channel consists of four member interfaces: Gi1/0/12, Gi1/0/13, Gi2/0/12, and Gi2/0/13.

Regarding your question about the bandwidth distribution across VLANs in the port channel, the bandwidth is not split equally between VLANs by default. The port channel aggregates the bandwidth of the member interfaces, but the distribution of traffic among VLANs depends on the traffic patterns and the load balancing algorithm configured on the switch.

Load balancing algorithms determine how traffic is distributed across the member interfaces of the port channel. Common load balancing methods include source IP, destination IP, source MAC, destination MAC, and others. The specific load balancing algorithm being used in your setup is not mentioned in the provided information.

LockStock

[Joseph W. Doherty]

As the other posters have already described, 1) a single flow will only obtain the bandwidth of a single link, and 2), distribution across an Etherchannel depends on the LB algorithm being used and the attributes of the actual traffic.  For the latter, worst case all flows are directed to only or same link.  Best case, you have equal distribution.  BTW, for last case, equal distribution very seldom achieved short term, but might be seen as a long term average.

However, one point often overlooked, without lots of flows, it's difficult to achieve all your aggregate bandwidth.  Why?

Consider 4 gig flows being directed to an Etherchannel of 4 gig links.

1st flow has a 100% chance of obtaining its own link.

2nd flow has 75% chance obtaining its own link but a 25% chance sharing the 1st flow's link.

3rd flow has same odds as 2nd flow if 2nd flow sharing link with 1st flow, otherwise it has a 50/50 chance of obtaining its own link, or sharing a link with either the 1st or 2nd flows.

Hopefully you can see odds for 4th flow are worst yet.

In aggregate, for 4 flows, you'll likely only obtain, on average, half the possible bandwidth.

Your OP did question why you might only be obtaining half your aggregate bandwidth, correct?