Content area
Full Text
IP networks were originally designed to be fairly simple. There's a source and a destination address, and the network devices use this information to perform some fancy calculations--and magically, things connect.
But as the internet has grown and more endpoints have been connected, networking has become a black magic. Since it's impossible to give every device its own unique IP address, the clever folks at networking companies came up with an assortment of workarounds, such as being able to NAT (network address translation) non-routable, private addresses. And as we've added more dynamic environments, such as private and public cloud, defining policy based on addresses or ranges has become unsustainable.
RELATED: Auto parts supplier has big plans for its nascent IoT effort
For example, a business may have thousands of devices and only a few public IP addresses. At this point, the business has a couple of options. The first is to only connect the number of devices for which the business has available addresses. This certainly seems impractical.
The second option is to address all the devices with private, non-routable addresses such as 192.168.X.X and then perform NAT when broadcasting to the internet. But multi-NAT environments, as well as Carrier Grade NAT, make it impossible to punch through and connect distributed systems. IP address limitations further restrict what can be done with regards to critical functions, such as cloud peering preventing the easy movement of workloads between VPCs, as well as between clouds.