cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
25806
Views
0
Helpful
9
Replies

IP Subnetting best practice

netsquant
Level 1
Level 1

Can some provide me documentation on splitting a class C network. I can't think of any good reason other than saving IPs but when your part of an organization that has an entire class B (10.29.0.0/16) that isn't a good enough reason for me. I know at least three reasons its not good, vlanning,Subnetting,dhcp. .........another words to many things to worry about. Human error very likely.

9 Replies 9

Reza Sharifi
Hall of Fame
Hall of Fame

tobyarnett
Level 1
Level 1

The largest reason to to break up your subnet is to conserve IP addresses. The other reason is to keep your broadcast domain to a controllable size. You should not have anything larger than a /22 network for any LAN block. Segmenting smaller than a /24 is often used when you need special requirements for the devices. Take for instance you wanted to connect in 5 network devices as part of a single subnet. You wouldn't use a /24 for that or you just wasted 250 IP's. So you can use a /29 instead and this will allow you to create many /30's subnets for point-to-point links. These seperate networks also makes it easier for you to create access lists on your routers or firewalls, writing policy map statements, configuring networks in your routing protocol configurations, and various other reasons. The list for why to break up a subnet is really way to long to list. I am not sure there is any specific document that lists all of these reasons, but you can get this information from pretty much any networking resource you find on the Internet.

The post above is a great site to use break up any subnet into smaller or larger blocks depending on what ever use you determine.

-Toby

Sent from Cisco Technical Support Android App

-Toby


Please don't forget to rate any helpful post.

_____________________________________
There are no great limits to growth because there are no limits of human intelligence, imagination, and wonder.
- Ronald Reagan

It was recommeneded to me to take a /24 network and split it up.  the first have for management and the second for production.  I think this is a huge no no.  I was told its to save IP's.  I don't understand that if we have an entire /16 network that we can carve out another /24.  One entirely for production and the other for managment.  I'm looking for some documentation stating this is best practice unless there is a shortage of IP's.

Sakun Sharma
Level 1
Level 1

Subnetting helps you not only save your IP's but also add security layer to your network. It divides your broadcast domains increasing your network performance.

Now if your whole network is one network without subnetting, 1 single broadcast packet will reach to all the devices connected (computer, printer, AP, Mobile, IP Phone, etc) in your network. Similarly, if any rougue device generates so many broadcast, broadcast storm will occur in your network and whole network will be down.

But if you are subnetting, broacast will remain within broadcast domain (subnetted network) only, it wont span whole network.

If whole network is in one subnet, all your devices can reach other devices in network, so your critical servers can be attacked easily. As we know that 80% of attacks are from within an organization, so its not safe to keep your complete network in one subnet. If you are using access-list also in one subnet, there will be large number of access-list entries to protect your network, further reducing your Inter-network switching device performance and network performance. But in case of subnetted network, you have more control our your network security and number of entries in access-list will also be less.

So using subnetting you can better organize your organization network, less complex, more secure, flexible and scaleable. So I suggest to go with subnetting and use of VLAN's.

--

Regards

Sakun Sharma

It was recommeneded to me to take a /24 network and split it up.  the  first have for management and the second for production.  I think this  is a huge no no.  I was told its to save IP's.  I don't understand that  if we have an entire /16 network that we can carve out another /24.  One  entirely for production and the other for managment.  I'm looking for  some documentation stating this is best practice unless there is a  shortage of IP's.

We use subnets not only to save IP's address, but also to tackle some issues at Layer 2 as well as Layer 3.

We should allocate the block considering the requirement and future expansion, because re-desinging the network is very difficult and unmanaged network in very difficult to troubleshoot. Large block size also leads to security issues as well as scalability issues. So its better to select the block size very carefully.

Below are some useful links and some stuff.

Check out the below link:

http://www.cisco.com/en/US/docs/solutions/Verticals/EttF/ch3_EttF.html#wp1016570

Cisco IP Address Guide PDF

Here is some stuff, hope you find it useful: (source Cisco Press: ICND 1 & 2)

A poorly designed network has increased support costs, reduced service availability, security risks, and limited support for new applications and solutions. Less-than-optimal performance affects end users and access to central resources directly. Some of the issues that stem from a poorly designed network include the following:

  • Failure domains: One of the most important reasons to implement an effective network design is to minimize the extent of problems when they occur. When Layer 2 and Layer 3 boundaries are not clearly defined, failure in one network area can have a far-reaching effect.

  • Broadcast domains: Broadcasts exist in every network. Many applications and network operations require broadcasts to function properly; therefore, it is not possible to eliminate them completely. In the same way that avoiding failure domains involves clearly defining boundaries, broadcast domains should have clear boundaries and include an optimal number of devices to minimize the negative impact of broadcasts.

  • Large amount of unknown MAC unicast traffic: Cisco Catalyst switches limit unicast frame forwarding to ports that are associated with the specific unicast address. However, when frames arrive at a destination MAC address that is not recorded in the MAC table, they are flooded out of the switch ports in the same VLAN except for the port that received the frame. This behavior is called unknown MAC unicast flooding.Because this type of flooding causes excessive traffic on all the switch ports, network interface cards (NIC) must contend with a larger number of frames on the wire. When data is propagated on a wire for which it was not intended, security can be compromised.

  • Multicast traffic on ports where it is not intended: IP multicast is a technique that allows IP traffic to be propagated from one source to a multicast group that is identified by a single IP and MAC destination-group address pair. Similar to unicast flooding and broadcasting, multicast frames are flooded out all the switch ports. A proper design allows for the containment of multicast frames while allowing them to be functional.

  • Difficulty in management and support: A poorly designed network may be disorganized and poorly documented and lack easily identified traffic flows, which can make support, maintenance, and problem resolution time-consuming and arduous tasks.

  • Possible security vulnerabilities: A switched network that has been designed with little attention to security requirements at the access layer can compromise the integrity of the entire network.

--

Regards,

Sakun Sharma

There is no technical problem with taking a /24 and splitting it into 2 /25s. The only issue is the capability of the it support staff to properly configure routing, subnetting, gateway configuration, acls, dhcp and other services using a network address boundary they are possibly unfamiliar with.

Sent from Cisco Technical Support iPad App

Jeff - that was my point exactly.  There for if there is no reason in this case to try and save IP's it's not a good idea.

"It was recommeneded to me to take a /24 network and split it up.  the  first have for management and the second for production.  I think this  is a huge no no.  I was told its to save IP's.  I don't understand that  if we have an entire /16 network that we can carve out another /24.  One  entirely for production and the other for managment.  I'm looking for  some documentation stating this is best practice unless there is a  shortage of IP's."

Jeff's point above is made in regards to how those subnet boundaries function and how your support staff can manage those. Take for example if you have a /25 subnet for a server zone. Then you network gateway and subnet will be something different than the typical /24 block and that may cause confusion with some server admins. However proper communication can resolve those issues.

However to the point of your above comment was to conserve IP address, just because you have a /16 does not mean you should be wasteful of how you address server blocks, network management devices, user workstations ect..

If you have a network that consists of 2 distribution switches, 2 edge routers, 10 additional network device (misc routers and switches), and 2 firewalls, then you could conscievable only need 18 IP addresses for these devices to be connected. In that case is it really necessary to allocate 254 IP's to manage your network devices? I would think that a /25 or even a /26 would be more than sufficient as this would give you ample IP's for you immidiate as well as more than plenty for growth, and you will likely still have waste. You could then carve the rest of the IP's out into /30's (for point-to-point routed connections), or some other smaller networks for device connectiity.

Likewise you should allocate server blocks based off of the needs of your business. If your company works with virtualized servers that utilize many IP's then you may want to look at creating server blocks that are in the realms of /23's, but if your company relies on physical servers and you only have 2 or 3 of those, then again why not use a /25 IP block for these. This will again allow your current needs as well as future growth.

You should break up your /16 into logical blocks (maybe in /20 blocks). This can be used for creating a hierachry of IP addressing that is useful for summarized routes in your routing table.

If you are designing a network you need to take into the needs of your business currently as well as try and predict what will be needed in the future. Split the networks out as appropriate, summarize when possible, and certainly do not waste IP space just to be lazy or have the thought process that "we have a lot of IP's so why split up our blocks". I have worked in many environments where someone that that was a great idea 10 years ago, and trying to fix that mess is a nightmare in a production environment.

-Toby

Please don't forget to rate any helpful post.

_____________________________________
There are no great limits to growth because there are no limits of human intelligence, imagination, and wonder.
- Ronald Reagan

-Toby


Please don't forget to rate any helpful post.

_____________________________________
There are no great limits to growth because there are no limits of human intelligence, imagination, and wonder.
- Ronald Reagan
Getting Started

Find answers to your questions by entering keywords or phrases in the Search bar above. New here? Use these resources to familiarize yourself with the community:

Review Cisco Networking products for a $25 gift card