As one of the first advisors brought in by Arrcus and co-chair of the IPv6 Operations Working Group, I am now continuing a long career in working with the protocols that make up the fundamental workings of the Internet. In fact, I’ve been involved with the IETF since 1989, including a stint as chairman. In that time, I’ve seen and participated in many of the technical and organizational advances that form the foundation of today’s Internet.
Today, the hot topic in protocols is IPv6 and how it’s now being adopted by many companies as a replacement rather than an extension of IPv4.
IPv6: now a “first-class citizen”
As IPv6 first became available, it was not necessarily clear how it would be widely adopted. This has resulted in an environment where IPv6 is widely used alongside IPv4. This currently causes a number of headaches for individuals who must manage networks involving both protocols, in part because vendors themselves continue to experiment in an effort to find the best way to adopt IPv6 on a wider scale. As a result, IPv6 has remained a difficult-to-liberate second class citizen among network protocols.
I have been a key technical advisor to Arrcus since its inception, and one of our primary focuses has been elevating the status of IPv6 to that of a first-class citizen amongst network protocols. IPv6 has now advanced beyond the early adopter stage, and is now moving into general deployment, all of which are strong indicators of our success. In fact, major networking organizations are beginning to take big steps toward regularizing IPv6. The UK-based hosting ISP Mythic Beasts wants to transition to IPv6 only service in the near future, and more and more networks are moving in the same direction, particularly for data center applications.
The abandonment of IPv4 in favor of IPv6 was never a foregone conclusion, and it has only been recently that the major tech firms including Google, Facebook, and so on have moved in the direction of actually replacing IPv4 with IPv6. When this becomes an industry standard, IPv6 will be a first-class citizen from the start. Why do these major companies want to go to an IPv6-only world? Simply, it makes their operations and network management a lot easier than having a dual environment. Because of the address crunch on IPv4, today’s IPv4 data centers are becoming increasingly complex with constructs including translation layers making problem diagnosis and many other functions more complex.
IPv6 is just one component in the total formula Arrcus is adopting to modernize data center management however. The other major component in this formula is tackling the massive complexity involved in managing data centers with multiple different kinds of hardware running on the same network.
Arrcus’s policy-based data center tools
The Arrcus team has been working on a technology for use in data centers that will run on just about any platform. This technology will consistently and reliably deliver services across that diverse range of hardware on the network. This offers a major advantage over the major network players including Oracle. In the model followed by these organizations, every hardware product is different in this regard is configured differently, and has different capabilities. This puts an enormous strain on network engineers tasked with maintaining networks incorporating diverse hardware. Of course, it’s often this diverse “white box” hardware that is the cheapest. In order to unlock those cost advantages though, firms will need to figure out a way to solve the complexity problem.
Over the course of my career, I’ve learned that this complexity is a hindrance, but it’s also quite unnecessary. Arrcus is developing a platform that offers consistency—with a single change you can update settings on all the machines on a network, and it simply works. At the core of Arrcus’s approach is a policy-based system that differs substantially from the model deployed by the big-name networking companies.
Central to data center technology nowadays is the concept of applying a single policy though one platform and subsequently applying it throughout the data center. Arrcus is bringing that policy-based approach to data centers. With Arrcus, rather than going to each individual device and configuring it manually, engineers can set one unified policy and have it implemented consistently and automatically across machine types.
Fred Baker is an Arrcus advisor with over 30 years experience working with Internet protocols. He has been involved in the IETF since 1989, including a stint as chairman. In that time, he’s participated in many of the technical and organizational advances that form the foundation of today’s Internet including adoption of the IPv6 protocol. Today, he serves as co-chair of the IPv6 Operations Working Group.