Once upon a time (Well, actually several different times, depending on how you keep score), there was a fiber "glut" - an excess of fiber carrying capacity. Too much unlit capacity on long-haul routes back in the late 1990s lead to a 2001 market correction among service providers and equipment manufacturers buying into a myth of an exponentially growing Internet with infinite growth potential. I don't think we'll go through another fiber building boom anytime soon, but there are a lot of bits and pieces to indicate more glass will be put to work in long-haul and short haul uses in the years to come.

Over the past couple of months, I've gotten an earful about small cells. AT&T and Verizon have both talked up small cells as the next big thing, with a senior AT&T technology rep describing all sorts sub-cell sizes, including femtocells, picocells, microcells, and metrocells. Small cell technology is being driven by numerous factors, including the need to provide phone coverage within a building and campus environments, "dense" coverage in public areas such as stadiums, and the need for more/better coverage for new LTE networks.

The WiFi Alliance was spinning its portfolio of technologies as a better, more cost-effective solution to LTE networks at CTIA last month, hauling out a white paper it commissioned saying the per-bit TCO cost for Wi-Fi is about half that of LTE. I'm skeptical of such claims, given that it takes years to establish standards in the Wi-Fi industry and that carriers have paid for and can cleanly manage RF spectrum they "own."  WiFi can be killed by the simple mechanism of turning on an old microwave oven.

Regardless, all these little itty-bitty cells will need backhaul/broadband connectivity.  Fiber will be needed because of speeds necessary to feed small cells and to more quickly move information between data users and networks to ensure bottlenecks don't occur because someone wants to replay the last episode of "Game of Thrones" on his iPad during a lunch break. Similarly, aggregating small cells will require still more fiber.

Cable companies have been mostly fiber for a long time, using the medium for long-haul networks down to the neighborhood, then distributing via Ye Olde Coax cable for the final "mile." Fiber also serves as their preferred method to deliver broadband services to their business customers. DOCSIS 3.1, the cable industry's latest and greatest standard, promises the ability to deliver up to 10 Gbps downstream and 1 Gbps upstream.

Not to be outgunned, Verizon is testing "micro-trenching" or "saw-cutting" in New York City to install last mile cable - a technique that appears to promoted by Google as early as 2010. Instead of inflicting whole scale destruction to roads, sidewalks, and other concrete with backhoes, a thin, almost surgical trench is cut just wide enough to drop in fiber, then filled back in.

It's more than the last mile that is getting buffed up. Long-distance, higher-speed fiber projects continue to cook along the globe. Verizon has invested in 100 Gbps routes - higher speeds at this time, not new fiber - across the United States and within Europe. The company has a 100 Gbps European ring of more than 2600 kilometers.

In Africa, there are numerous fiber projects both around the continent and within individual countries. Forward-thinking companies such as IBM and Orange are investing in the areas to drive the next wave of ICT growth over the next decade as markets in the U.S., Europe, and Asia remain relatively stable (and slow).

The biggest gamble for new fiber deployments will be over the Arctic Circle. A pair of companies plans to run projects through the legendary Northwest Passage to connect Japan to the United Kingdom, cutting latency for high-frequency stock traders and providing redundancy to other cross-ocean links. Global warming and icebreakers could make top-of-the-world fiber possible.

All the new fiber being built out for cellular, cable, and telecommunications providers will also require upstream improvements to more quickly move data throughout the network. In high capacity and/or longer distance routes, older fiber is going to end up being supplemented and/or replaced by the latest-and-greatest optical product with less loss and more capability to get maxed with the highest speeds possible.

Enterprises will have more options for connecting data centers at faster speeds from all these projects in the works, regardless of where corporate offices are located. Purchasing dark fiber for long-haul connectivity will likely become more affordable as competition increases and new projects come on line.

Life gets more interesting when you look at the metro and last-mile requirements. If you're in the right city and close enough to the "right" buildings, competitive deals for fiber connectivity can be had. Techniques like microtrenching can bring options that would have been previously cost-prohibitive, so if you've been eyeing a campus or short-haul fiber project, now is the time to update you background information.