Over the past couple of decades, the Computer networking industry has seen many trends come and go. Looking at how things stand today, the core skill sets of Routing and Switching look likely to become dated as the number of experts proliferate the market and the task of routing the packets through the Internet has become a commodity for so many deployments. Fundamentally, the core job of routing remains the same, from the perspective of a device. For it, a packet (or data) comes in and it needs to make a decision on where to send it out to. The USP for any device resides in the speed with which it can make this decision and the speed with which the packet can be forwarded out another interface, towards its ultimate destination. As with so many other technology industries that reaped the benefits of Moore’s law, it was the same with the High Performance Networking industry. Shrinking space for the transistors used for switching meant that so many more operations could be performed for the same real estate on the device.
The quantum of data storage has increased exponentially too. If we look at the changes around us even in the past 10 years, it is nothing short of extraordinary. For me, school and college days consisted of 5 and a quarter inch floppy disk drives which could store, maybe a few KiloBytes of data. That, in itself, was a novelty. In what appears to be no time at all, we transitioned to 3 and a half inch floppy drives which had maybe double the capacity of the bigger version. Then, we moved to CDs, USBs and today, the cloud is the place for storage. As in every stage, I tend to think that this is the ultimate method and can we really go bigger than this? My answer arrives in a only a few months. Data transfer has had to keep pace with this remarkable pace of development. Of course, as data transfer guys, we do lag behind. Storage becomes important only if the transfer is not quick enough. Come to think of it, if there were a world where the data transfer is to the tune of a few zillion bytes per second, what would be the need for data storage at end locations? I’ll leave the thoughts around the possibilities of data storage and data transfer for a later blog.
Trends in the Networking industry space point to higher speeds via technological advances. We have moved from 1 Mbps speed for Ethernet to 100 Gbps speeds. Standards for 400 Gbps and 1Tbps speeds are in progress. These will continue. Advances in the data switching capability will lead to new architectures on the devices themselves, replicating the methods used at a large scale deployment, to a micro scale at a transistor and chip design level. Consistency in the design architectures at every level of abstraction from a chip design, bring up, to a router/ switch, to a deployment scenario with hundreds of routers will lead to speeds and capabilities, outside of our imagination of today. In all this, we will reuse what has been done, build on it and (re)create fantastic new developments in all areas.
For the immediate future, the hype engine is over time in promoting Artificial Intelligence. Every company wants a piece of that pie that is AI. How will data transfer and communication evolve with AI? I am absolutely certain that even today, there are hundreds of university Professors that are deliberating and working on this particular trend. AI and Machine Learning will lead to even more interest in data transfer as the volume of information acquired will simply explode. I see Blockchain as another huge disrupting factor. How will this disrupt Networking? I do not know. I do know that it will lead to even new technologies and changes in the way we think of High Performance Networks.
The third major trend will be with Security. How do we ensure that only the people or systems that are authorized, have access to data?
Companies in the domain of High Performance Networking and so many others will need to realign and focus on these three areas to remain competitive and maybe even take the lead. You can bet that there will be startups and new initiatives galore aiming to jump on the bandwagon and make it big. It will be fun and interesting to see who comes out on top in the next 10 years.
Mithun's Musings 