Issue link: https://cp.revolio.com/i/661564
THE DATA CENTER JOURNAL | 11 www.datacenterjournal.com resPonsiBLe groWth Perhaps the biggest challenge facing today's data centers is energy. If existing trends are allowed to continue, data center energy consumption will surpass the cur- rent production of the whole planet. Envi- ronmental goals and legislation are driving scrutiny of carbon-dioxide footprints in every business vertical, and custom- ers everywhere are seriously considering corporate environmental behavior when choosing where to take their business. But for energy-intensive data centers, this problem is even more acute. Even though the main drivers are as much economical as ecological, getting serious about going green is the only way forward. It's going to take more than shaving off a few watts here and there to make a serious impact. Overall efficiency needs to be improved through architecture en- hancements that remove processing layers and reduce redundancy. For example, one straightforward way to increase Ethernet efficiency is by using data rates that are integer multiples of each other; doing so means that multiplexing rates up and down requires less processing. is fact explains the recent decision of standards bodies to forgo jumping to 1TbE and to instead fill in the data-rate gaps between 10GbE, 100GbE and 400GbE with 25GbE, 50GbE and 200GbE. In addition to reducing power consumption and the associated CO2 footprint, other aspects such as logistics, resource management and the supply chain must also be considered. Data center operators need to employ the concept of circular economy to, among other things, extend product lifetime and optimize re- cycling. Only with a holistic sustainability approach, involving a radical overhaul of logistics and selecting equipment suppliers with products that are proven to be energy efficient, will data centers be able to com- pete and survive. is the goLD rush Waning? But because the most popular topic this year in data centers and the cloud is how to keep up with growth, there is a quiet counterculture asking the explo- sive question of whether profitability has peaked. Like nuclear power plants, high- way systems, undersea "chunnels" and so on, data centers and global cloud services are a huge infrastructure investment. And to justify these long-term investments, there must be sufficient long-term promise of profitability at the needed margins. Of- ten, infrastructure building peaks on hype and a fresh influx of investor capital, just as competition forces business models under- water. In several high-profile cases, large service providers have given up on cloud services, primarily because the profitability did not justify the required investment level. As an analogy, consider the case of a transportation link, highway, train tracks, canal and so on. e type doesn't matter. You are only going to spend resources to update it if the cargo flowing over it is making enough money. Similarly, unless cloud services over the top of the data center and global network make sufficient revenue, there is little reason to continue to invest in the underlying infrastructure of the cloud. And currently, the content owners themselves are making money off the cloud. Consequently, service provid- ers are of the opinion that if the content owners are making the money, let content owners build the cloud. And build they have, as Internet content providers are now constructing transport networks that will dwarf those of service providers. CLouDs on the horizon So where does that leave data centers? Although the future is always uncertain, several trends point the way. Power will continue to be important and drive most decisions. Exponential growth will continue, though GBF (get big fast) will be replaced by more-measured, sustainable growth with a greater focus on environmental impact. Network hard- ware and soware will be more open. e networks themselves will be run and monitored by bots. e cold war between hackers and defenders will go on. And like most everything else in this world, money, profits and gross margins will continue to drive decision making. n about the author: Jim Theodoras is vice president of global business development at ADVA Optical Networking, advising data center operators, content providers and enterprises on how to best interconnect their data centers. He has over 30 years of industry experience in electronics and optics, spanning a wide range of topics. He holds 20 patents in the field of telecommunications and is a frequent contributor to industry publications. But because the most popular topic this year in data centers and the cloud is how to keep up with growth,there is a quiet counterculture asking the explosive question of whether profitability has peaked. Like nuclear power plants, highway systems, undersea "chunnels" and so on, data centers and global cloud services are a huge infrastructure investment. And to justify these long-term investments, there must be sufficient long-term promise of profitability at the needed margins.