Cloud backbones are continually growing and evolving to accommodate more and more services. I am reluctant to quote any figures here, due to their fluid and somewhat subjective nature, but it’s widely agreed they consist of massive numbers when it comes to regional coverage, number of points of presence (POPs), miles of fibre deployed, and gigabytes of bandwidth offered.
In a previous post, I discussed how ‘the cloud’ has redefined the need for enterprises to host their own network infrastructure, and with it, made the need for Regional Internet Registries (RIRs) to promote IPv6 to such organizations redundant.
I want to continue exploring this theme of how the cloud is disrupting Internet management practices/roles due to ongoing significant growth of cloud backbones, the sharp rise in cloud-based infrastructure assets and virtualized resource deployments, and the bulk migrations of enterprise data centres carried out by cloud providers. However, this time, I want to question if RIRs are ready for this changing landscape, and if they foresee their responsibilities being somewhat limited in days to come?
Is ‘the cloud’ better than ‘the Internet’?
If you’re like me, I always thought there was only one Internet in this world. But in this cloud era there now seems to be two distinct kinds of Internet — a ‘public Internet’ and a ‘non-public Internet’.
It seems that cloud operators don’t like to be theoretically included in the Internet even if in reality they are part of it, perhaps due to business/marketing reasons.
Non-public doesn’t necessarily mean private, as these cloud resources are regularly accessed by the public over the Internet. Instead the name refers to the cloud operators’ own flavours of Internet (global networks), which they claim, are far better in performance and security than the regular public one.
As a cloud architect, I’ve learned that cloud providers make every effort to make their networks as fast, efficient and secure as possible. After all, there are multiple cloud services they are all competing against.
Aside from their marketed differences though, they all want to keep their customer’s IP traffic within their global network (cloud) for all service/infrastructure/resource needs. This is where I see the true competitive advantage over the public Internet as it gives them far greater control and more specific metrics on their network’s ability and health. Put simply, cloud networks are way better designed than ‘the Internet’, which is a Frankenstein of protocols and extensions.
For this reason, it seems we are heading towards a ‘proprietary Internet’ era where cloud operators will be operating and managing their own brands of Internet. And the Internet they now refer to as the public Internet will perhaps be limited to edge networks (last mile or first mile) for accessing various global resources from cloud networks.
Are RIRs ready for this changing scenario?
In my opinion, RIRs will continue to distribute IP addresses, promote IPv6, assign Autonomous Systems Numbers, recommend routing optimization and security best practices, and raise awareness among Internet users, all of which they are already doing nicely.
However, when it comes to how IPs are delegated and used on cloud backbones, RIRs may not be able to accomplish one of their key responsibilities — keeping their whois registries up to date — due to them not necessarily being public.
As vendors justify their business based on the shortcomings in the public Internet, we all are keen to see how much influence RIRs will still have over the entire Internet.
Azhar Khuwaja is a Telecom/IT Trainer with over 20 years of industry and training experience.
The views expressed by the authors of this blog are their own and do not necessarily reflect the views of APNIC. Please note a Code of Conduct applies to this blog.
Excellent Wite-up.