Supernet: What it is and what it is for?

From now on, let us add “Supernet” into our vocabulary. Learn more!

A set of computers and/or computer equipment connected to each other, and that can exchange data and information, all of those make up a network. The Internet is the network of networks. We could even think that the Internet is the “Supernet”, but we have to tread carefully when using frequently debated terms in computing… For that reason, today, we bring to the fore the term “Supernet” (or Supernetwork), of course always taking an approach from a monitoring perspective.

* Warning: what I write below is my way of looking at things from a practical and sincere point of view. This article only endorses me and, in any case, this entry should be read taking into account the learning approach, it does not intend to be official in any way. That said, let’s start from the basics, which is not the same as starting from scratch.

Terms: Supernet and Supernetting

If we have a network and we buy a new computer, we say that “we add it to a network.” If we have a supernet, then “we add it to a supernet.” We even have a specific verb for it. It is very common to use the term Supernetting; however the following terms are also valid (but less used):

• Prefix aggregation
• Route aggregation
• Route summarization

If we get even more specific, technically we will find differences but for the purposes of this post we will deal with it the same way… Do you think it is daring on my part? Well, there is more!

Request for comments

Although we can go a lot further back in time, the Internet was born in the United States of America, originally called Arpanet, in the late 1960s. A technological predecessor could be the landline from which many of the concepts are born, used when planning the “network of networks”. In fact, the wiring itself, the colors that identify the pairs, are very similar, at a physical or hardware level. This includes the similarities in switched connections (or circuit switching). But, obviously, the Internet and data transport in a digital way ended up completely absorbing telephony.

But the Internet needed more than the physical and conceptual foundation or sustenance of the great American telephone companies. Moreover, October 1969 is marked as the birth of the Internet since the first connection between two computers was made… And it was simply that, since it was not yet a common computer network.

The Internet was born, in my opinion, when pioneer Dr. Steve D. Crocker published the first issue of Request for Comment (RFC) on April 7, 1969. In Issue 6 (RFC 6), Steve Crocker recounts his conversation with Bob Kahn about code conversion for data exchange. RFC 11 publishes the connection implementation in the FAT operating system (yes, that’s what it was called), and I fervently believe that this, published in August 1969, is what enabled the feat performed in October of that same year.

Based on this knowledge base, the RFCs were born: gathering a group of people in their twenties who moved among different universities sharing knowledge and cementing concepts, something that we now do by email… In fact, RFC 733 (1977) outlines this technology and the standard for the email is published in RFC 822 (year 1982).

RFCs grew decade after decade: in 1992, RFC 1338 “Supernetting: an Address Assignment and Aggregation Strategy” was published for information purposes. Yes, at first the supernet was just a mere advertisement, not a protocol, and not even a standard.

Just the following year, in 1993, RFC 1518 “breaks” the paradigm of networks by classes. While class A networks allow millions of IP addresses, the next step – class B networks – only allowed 65 thousand IP addresses: between the two of them the “waste” of IP addresses is very high.

For that reason, the Classless Inter-Domain Routing (better known as CIDR) was born, which is an extension of the original IPv4 addressing system that allows more efficient address allocation. The original class-based method used fixed fields for network identifiers, which was wasteful as I said earlier: most organizations that are assigned those addresses (class A and class B networks) never intended to put so many devices on the Internet.

As additional information, this is the origin of CIDR notation, the suffix that accompanies an IP address (there are 32 bits in an IPv4 address, four octets separated by periods) and that allows describing or narrowing down a range of them. For example, for /20 it allows 4096 IP addresses, for /21 2048 IP addresses and so on, as well as all the way around (all powers of base 2, this is important for a supernet as we will see later). All these numbers can be obtained using the IP address calculator included in Pandora FMS. You may also find many of these calculators online, each with its own style, shapes and colors to present the same data.

Flexible like Pandora FMS

CIDR thus changed the fixed fields to variable length fields and this allowed to assign IP addresses better, and in a more refined way. CIDR IP addresses include a number that indicates how the address is divided between networks and hosts.

For example, in the CIDR address 201.249.0.0/19 the /19 indicates that the first 19 bits are used for the identification of the network and the remaining 13 are used for host identification.

The main purpose of a supernet is to decrease the size of the route table of routers. For example, instead of a router having 8 individual routes, it may have a single route aggregated from these 8 individual routes. This saves memory and processing resources on the routing devices, thus requiring less space to store their route table and less processing power to search the route table. It also provides stability in networks because fluctuations can be isolated, that is, in one part of the network they do not spread to all parts of the network.

Supernetting and Pandora FMS

From Pandora FMS version NG 731 IPAM was included (abbreviation of Internet Protocol Address Management) which allows to manage, discover, diagnose and monitor hundreds of IP addresses.

Within this feature, the supernet, subnets and even virtual private networks (VLAN) are included, all integrated, with the option to export data in CSV files. Unlike creating VLANs, we can only create supernets manually using IPAM. For that, you have to configure, with the necessary parameters, each of the supernets that you want to have, and later add networks already managed with IPAM that may belong to a VLAN. Although it is a manual process from version NG 758, it includes the ability to quickly addy our data from files in CSV format.

To finish off this post, let’s see what the rules that operate a supernet are.

Supernet Rules

Apart from good practices in network configuration, the established rules must always be followed and enforced to avoid chaos reigning.
The rules for creating supernets are as follows:

• Networks must be contiguous or sequential.
• The number of networks to add must be a multiple of two or “base two”.
• And the rule that is somewhat more complicated: compare the value of the first octet not common from the first block of IP addresses (the smallest) of the list of networks to add against the number of networks to add (see previous point). The value of the first non-common octet must be zero or a multiple of the number of networks to be added.