I was recently talking to a good friend who is in the process of looking to buy a house. He and his wife live alone, with no children or other relatives to require massive amounts of space. Nevertheless, they were looking at houses that were quite large.
Knowing that my friend is not generally ostentatious in his preferences, I asked him why he was looking for so large a house, and he responded that he and his wife wanted plenty of space to entertain and have plenty of bedrooms for guests.
Curious, I asked how often they hosted parties. “Not very often,” he replied. How often do they have overnight guests? Almost never.He noted that they might do both of these things more frequently if they had the space, but over the course of the conversation it became clear that my friend may have been taking on far more house than he reasonably needed on the off chance that it may fulfill its usefulness a few times a year.
How often, in the network expansion planning and deployment phase, have you seen this same pattern?
Very often, teams within a carrier's organizational structure push for massive infrastructure expansions that don't match the realities of current and projected network use. I've heard numerous accounts of IP Transport Engineering teams receiving extraneous bandwidth expansion requests from the Optical Core and Radio Network teams. Since these teams don't have visibility into the actual IP traffic domain, they have to rely on optical measurements and traditional mobile infrastructure measurements based in Erlangs at the RAN to get a feel for network traffic.
Based on these kinds of measurements, teams believed that their domains were getting saturated, so they requested more bandwidth. The problem is, their requests were far greater than real network traffic patterns could justify. In one account, the Core team requested upgrades from 10Gb to 40Gb when the actual traffic was only clocking in at 8Gb. Meanwhile, the RAN requested a capacity expansion to 20 Mbit backhaul links when they were, in fact, only using 3 Mbits.
Now, more capacity isn't always a bad thing, but when requests are this far away from reality, it's important to consider the tangible costs that these network expansions entail for wireless carriers, in terms of both CAPEX and OPEX. Take the backhaul expansion costs as an example: If the carrier owns the network, you're talking about a massive CAPEX spend (millions of dollars in equipment and potentially additional microwave spectrum) to build additional capacity that is unlikely to be used in the near future. However, the costs are even more apparent when the carriers don't own the backhaul connections, and have to instead lease the connections and staff a team to monitor them, resulting in huge amounts of waste in unused OPEX-impacting network capacity.
There is, naturally, a third way. Mobile carriers can avoid the unnecessary expense of deploying and managing an overbuilt backhaul network, while also avoiding the QoS headaches associated with a network that can't handle the bandwidth its end-users demand. The key to this elusive “right-sizing” all depends on network visibility.
With the right tools, transport departments can validate the need and accuracy of network expansion requests by Optical Core, RAN Access, and other internal departments. In addition, this same type of network and traffic visibility can help assure high levels of service performance after the expanded network capacity is deployed.
Furthermore, it's extremely useful, from an operational standpoint, if this level of network utilization visibility can be shared across departments and domains, rather than being confined to a single department. That way, users in Core and RAN don't have to submit requests based on overestimates or underestimates of network use, only to have their requests overridden by departments with more strategic insights into true traffic patterns. It's a time-saver, and it promotes overall efficiency.
So the question for carriers is this: Do you have a clear sense of how much “room” you need? Or are you prematurely investing all the money on a few over-commissioned links rather than distributing the investment across the entire network in an efficient manner. Without clear cross-domain network visibility, it's a guessing game.
This post is a part of our Maximizing the Ethernet Mobile Backhaul Opportunity Expert Blog Series. We invite you to read other recent installments in this ongoing series:
View part 1: The Pipeline Ethernet article: Harnessing the Potential of Ethernet Backhaul by Juan Prieto, Product Marketing Manager — Mobile Solutions, InfoVista
View part 2: Harnessing the Potential of Ethernet Backhaul by Juan Prieto, Product Marketing Manager — Mobile Solutions, InfoVista