There was a mistake in my “Simplified IP Addressing” article that appeared in the January issue. The problem is that, with the current IP methodology, the number of subnets is given by 2<+>n<+> rather than 2<+>n-2<+>. So for the Class B subnetted to a Class C, we get 2<+>8<+> = 256 subnets. The 2<+>n-2<+> is correct for the number of hosts.
Moving on to “Subnetting on Non-Byte Boundaries”, the same 2<+>n<+> formula applies to the number of subnets. Hence, for the case I took, viz. borrowing three bits for subnetworks, we obtain 2<+>3<+> = 8 subnetworks, not 6. Accordingly, two more subnets should be added to Table 6, that is:
Subnetworks for Class C Network
With these two additions, Table 6 should now show 8 subnetworks. Likewise, Table 7 should be modified to show the 8 subnetworks rather than 6, with 0 being the first and 224 being the last.
Analysis of 256 Values of Last Byte
Finally, the analysis that follows Table 7 should be changed accordingly. Simply put, there are 8 subnetworks with 30 host addresses each, so we have 8*30 = 240 addresses available. Now, to determine the number of lost addresses, use 256-240 = 16. So, we lose only 16 addresses. This is much better than the old system.
—Prof. Gene Hector, geneh@acc.netsnwebs.com
