I our current environment of CM 2012, we have created boundary via IP address range; however we observed and lot of issues like boundary overlap with the same, ee have around 2000 ip based boundaries.

Please suggest whether IP subnet based boundary are better and easily manageable.

 

See http://blog.configmgrftw.com/ip-subnet-boundaries-still-evil/

Ultimately, it's a difficult and potentially subjective question to answer.

My first question for you is whether those 2000 boundaries are aggregateable or not? In other words, are there consecutive ranges that are within the same boundary group?

For example, let's say you have boundaries for 192.168.117.0-192.168.117.255 and 192.168.118.0-192.168.118.255 in the same boundary group. Then these are aggregateable into 192.168.117.0-192.168.118.255.

In general, a well laid out network will have many consecutive ranges like this to reduce the size and increase the efficiency of the routing tables. The concept is identical here and is easy to implement although it will take a lot of staring at numbers. In the organizations where I have gone through the exercise of aggregating ranges, I usually get a reduction of 30-40% of their boundaries.

yes Jason there a lot of consecutive IP ranges in the environment; however the concern going with IP address based boundaries are the boundary overlaps, which recently caused escalations.

I went through the link http://blog.configmgrftw.com/ip-subnet-boundaries-still-evil/ 

and what i understood is sccm client (IP add+ Subnet Mark) generates its Subnet ID, and this subnet ID checks the IP Subnet based boundaries Subnet ID and when Subnet ID matches the client look for the content on the DP added to that boundary or boundary group.

Now if the subnet mask for the boundary is different or incorrect, then client does not fall into a defined boundary (or boundary group).

So we need a have accurate subnet mask for correct reporting of clients.

Please correct if i understood correctly.

That's true. 
Which issues were caused by the overlap in your environment? Content was downloaded from the wrong DP?

What happened was  that both the IP based boundary which overlapped have common Distribution point.

Now what was explained is that the data circulated in the overlap boundaries causing latency and sccm client although having some other local DP looked for content from the above DP.

Although I still not 100% sure about the above. May be you can give your views.

Yeah, that doesn't make sense. Overlaps in boundaries within the same content location boundary group are harmless.

If the boundaries are going to be assigned to the same DP(s), then they should be in the same boundary group and by aggregating them, you reduce the number that must be used.

For you comment above about IP subnets, not sure why you are bringing that up as that has nothing to do with IP Address Ranges. Are you considering going with IP Subnet boundaries? If so, then yes, your comment is accurate. You must know the subnet mask of the *clients* to accurately determine the IP Subnet boundaries.

If you've properly aggregated and assigned your IP Address ranges to boundaries *and* your SQL server is not underpowered, then you should have no issues though.