We are in the process currently of helping a number of clients make major upgrades to networks, something we’ve done many times over the years. And this got me thinking about obsolescence and when and why we replace major electronics.
There are a couple of different kinds of obsolescence. First is physical obsolescence, which is when we replace things because they simply wear out. We do this all of the time with vehicles and hard assets but it’s rare with electronics. I can only think of a few times over the years we’ve helped people replace things electronics that were failing due to age. A few that come to mind are some T-carrier systems in the customer network that lasted for far more years than anybody expected.
A more common phenomenon is functional obsolescence where the electronics are not up to the task of handling newer needs. While this can happen with all kinds of electronics, the most common such upgrade has been replacing the electronics on fiber backbone or long-haul networks. There has been such a prolonged explosion in the amount of data our networks carry that it’s been common to overwhelm transport electronics.
In these cases we yank out fully functional electronics and replace them with something that can handle a lot more data. I would hope in the future that we will see a little less than this. One of the reasons we’ve needed these kinds of upgrades is that network engineers would not consider exponential bandwidth growth into their future projections. The naturally conservative nature of engineers didn’t let them to believe how much traffic would grow in just a few years after they build a network. But I finally see a lot of them getting this.
We also see technologies that are much more easily expandable. For instance, a lot of fiber electronics are now equipped with DWDM and other tools that allow for an upgrade on the electronics without a forklift upgrade. The network operator can light a few more lambdas of light and get a boost in throughput.
My least favorite form of obsolescence is vendor obsolescence where functional equipment is made obsolete when a vendor introduces a new generation of electronics and stops supporting the old generation. Far too many times this feels like nothing more than the vendors trying to force more sales onto their customers rather than looking out for the customer’s best interest.
This is not a new phenomenon and there was nobody better at this in the past than companies like Nortel and Lucent. They constantly pushed their customers to upgrade and were famous for cutting off support to older equipment while it was still functional. But the practice is still very much alive today.
Losing vendor support for electronics is a big deal to a network owner. It means you will no longer be able to buy a replacement for a card that goes bad unless you can find one on eBay. It means that the vendor won’t talk to you about any problems that crop up in your network.
The industry is now entering the second round of vendor obsolescence with FTTH electronics. Vendors cut off BPON and other first generation FTTH gear almost a decade ago and are now planning to do the same to GPON. I remember when BPON stopped being supported that every vendor of the next generation of equipment promised that the newer generation of electronics would be frontwards compatible – meaning that the ONTS and field electronics would work with future generations of core electronics. But as I always suspected this isn’t going to be the case and there is going to be another forklift from GPON to next generation of PON electronics.
The shame of this is the older PON equipment still works great. I have a few clients who have kept BPON working for a decade after it was supposedly obsolete by buying spares on eBay. Those networks are now finally becoming functionally obsolete as customers are using more data than the network can handle. But the equipment became functionally obsolete ten years after the equipment was declared as vendor obsolete. Most BPON electronics were well made and the ONTs and other field electronics have been chugging along a lot longer than the vendors wanted.
It’s not always easy to decide to keep operating equipment that the vendor stops supporting. But I’ve seen this done many times over the years and I can think of very few examples where this caused a major problem. It takes a little bravery to keep operating equipment without full vendor support, but management often chooses this option from the pragmatic perspective of economic reality. Most networks don’t make enough money to fund replacement all of the electronics every seven or ten years, and perhaps it is lack of money as much as anything that provides courage to network owners.