This Industry Viewpoint was contributed by David W Wang
As 2016 just kicked in, the news swirled that Verizon Communications may start to sell off their data centers for billions of cash in return. While such news is not really a surprise, since other major telcos like AT&T, CenturyLink, and Windstream have already done or considered doing so, it is surely dramatic. Only five or six years ago these telcos were rushing out to buy and acquire data centers and cloud computing resources into their pockets, but now they seem to want to kill these hens that once laid golden eggs. Of course the telcos shedding their data centers doesn’t necessarily mean data center is a declining business. On the contrary, data centers play pivotal roles in the next generation datacom, cloud computing, and data intensive wireless services. However, some serious flaws must have been troubling the current model of the data center business, hence why now the telcos are determined to deal with these pains. What would be the impact of this data center divestiture wave to the industry, market and customers?
If following the common thinking track, we may expect some tier-1 data center colo service providers like Equinix, Digital Realty or CoreSite stand out and grab the data centers on sale. After all, using Verizon Terremark as an example, its “crown jewel” assets with incumbent customer base there like NAP of the Americas in Miami Florida, NAP of the Capital Region in Culpepper, Virginia, and NAP of Amsterdam in Europe are all very popular and valuable. In my opinion, if Verizon acts quickly with the right offers, there might be a window for it to seize and unload the data centers with ease to some other colo or web service operators. But the window won’t be that big and still. Things can change and evolve fast for the data center industry and market.
Three major forces or issues may interact with and shape the real impact of this data center spin-off by the major telcos. First of all, the increasing pressure from the high cost of data center operations, in terms of power, air conditioning, hardware and software systems, and infrastructure management, etc. is threatening the sustainability for a sound business model. Some significant and innovative solutions in this regard are due. For example, major colo service providers are investing into renewable energies such as solar, hydraulic and wind power to offset the power consumption of their data center fleet. Major web service providers like Amazon or Facebook are using ODM (Original Design Manufacturer) to bypass brand vendors like HP or Dell for data center infrastructure purchasing and deployment. In this way, they can save 25-30% of the total hardware cost.
All these aggressive cost saving initiatives, however, would work out better over relatively new asset sites with new geographic selection, design and deployment which require less cost in innovation and re-engineering. The Verizon Terremark assets, on the other hand, are not the new breeds of data centers in its design, architecture and operation. As time goes, they may quickly depreciate due to technology evolution and business model changes.
Second, the new breeds of data center are getting smaller, commodity driven, and proximity to the customers - the so called edge data centers. Getting smaller is quite like the way the hotel business operates. A major hotel chain must find ways to offer more affordable lodging with the same core service quality to customers, such as a 5 star vs. a smaller 3.5 star hotel. Keep in mind this is different from the existing so called Tier 2 and Tier 4 data centers which carry quite some gap in service core functions. The new breed data center can still be a tier-3 or 4 level in capabilities but just smaller in scale and size. Is this feasible? Sure, advanced software system and virtualization technologies will help make all this possible.
Commodity driven refers to the popularity of basic cloud, storage and backup services on the bread and butter level. As more enterprises and technology service vendors moving and centralizing their IT operations to the cloud and data center colo space, not everything will require high end solutions. Quite many items will just fall under routine admin and maintenance needs that can be accommodated by basic commodity like services. Edge data centers, on the other hand, tend to get closer to the end users for smooth migration, shorter traffic backhaul, low network latency, and easy access for daily management. So while the super centralized tier 3 and 4 data centers remain popular, the new design and buildout of local and edge data centers may soon steal the thunder and take a lot of business away from the crown jewels and make the data center business more distributive in the future.
The third factor is the tradeoff between consolidation and decentralization. The telcos’ spinning off their data center assets may logically create some capacity and location surplus and duplication for the industry. So consolidation may be needed to handle this situation. Meanwhile, as stated above, another trend is for smaller, basic service and closer to end user data centers to quickly expand and grow. It sounds conflicting, doesn’t it? A potential middle path might be for colo service providers to dismantle those super level costly data centers and then diversify its existing tenants to other new breed and smaller edge data centers. This may make a win-win for the industry and market, because colo service providers can save cost and resources on one hand, and their clients will be better served with less service glitch and more care on the other hand.
All in all, this wave of data center spin-off by major telcos may not just go as another routine round of re-grouping and alignment for the industry and market, rather it may push for the arrival of Data Center 2.0 - the new revolution of data center service models, technologies, and practice.