#fabricpath

Back when the NASDAQ was flirting with 5000 and the bull market seemed to know no bounds, there was an innovative company called Palm that had put forth the notion of downloading web content as 'Web Clippings'. This was hailed as quite the revolution as it would open the door for web applications to be accessible by hand helds which in turn would synchronize with the 'Internet' once docked in a cradle hanging of a PC. 

Well, a decade later, the ubiquity of data networks and open standards such as WAP put an end to that notion. If there was a transitory technology it was that - it filled a gap for the few years that handhelds lacked native IP connectivity and the carriers were too busy protecting their voice networks to see over the horizon. 

Similarly, there is a potential that the new innovations in network engineering may diminish the necessity of dynamic DNS and large scale IP-MAC mapping management. When virtualization first arrived on the scene (circa ~2003ish), it became soon apparent that to overcome the IP changes that virtual machine movement requires, DNS infrastructure needs to be resilient, readily scalable and one that allows for management enmass of thousands of hosts. 

That spawned off a growth in appliances to address the need for data centers. Some of the notable vendors are Bluecat Networks and Infoblox amongst limited functionality being part of product lines from F5 Networks and Avaya (amongst others). This is becoming a necessity as IT operations seek to address smaller RTO windows for critical applications by automating the migration of workload across data centers. Of course, most network engineers would note that such an approach is a kludge at best of times and more of a hack at others. There are issues around TTL and having the DNS space stabilize after transference, requiring a level of tolerance of the applications while ensuring that the OS layers is optimized for such an approach.

Now with the adoption of IPv6 (more of a stimulus of change) and flatter network topologies through next generation protocols such as QFabrics, FabricPath etc., it is a matter of time whether there would be a need for dynamic DNS to the extent today. Though there will be certain use cases that would need it, the days of dedicated infrastructure to support it would wane in large shops. It would be easier to have the large network vendors (ergo Cisco, HP and Juniper) built it into their management tools rather than having a bolt on solution.