Why should a frame contain even number of octets?

[Joseph W. Doherty]

Because the spec says so.

As to why, I suspect because most almost all computers work with a "word" size, generally comprising a number of octets (bytes), also such "words" are always (?) powers of two.

[mad_mavula]

Can you explain it more.

[Joseph W. Doherty]

I possibly can.  What, specifically, don't you understand or need further explanation of?

[mad_mavula]

As per my knowledge word size is something has to do with processing data. But why the NIC identifies frames that has odd number of bytes (octets), as frames with alignment errors?

[Joseph W. Doherty]

"As per my knowledge word size is something has to do with processing data."

Yes and no.  True, "words" often contain data (BTW and what does a frame comprise, but data?) but (the no) some computers' instruction have "rules" with "words" too.  Actually, a computer's word is often, somehow, related to its data and processing register sizes.

"But why the NIC identifies frames that has odd number of bytes (octets), as frames with alignment errors?"

I would presume because an even number of octets was expected.  I.e. frame is not aligned with spec.

[mad_mavula]

Thank you for your support!!!