|
近年来，边缘计算的概念频频出现并迅速推广普及，边缘计算服务更靠近用户，能够大大减少在云中心模式站点下给应用程序所带来网络延迟、低带宽等影响。而边缘计算网络的搭建离不开本文提到的城市数据中心，城市数据中心选址在调研阶段就需要充分考虑到周边的市政设施情况、空气许可与湿地限制等环境因素、电力可用性等，只有充分考虑到这些不利因素并合理解决，才能推进城市数据中心项目稳步进行。
城市数据中心的选址必须考虑市政基础设施、空气许可、环境限制和电力服务等相关要素。
Site selection for urban data centers must consider municipal infrastructure, air permitting, environmental constraints and electrical service.
近年来，数据中心行业的建设规模出现了明显增长，尤其是在城市环境中接近人口中心的地带。对于数据中心的业主和开发商来说，接近人口中心的城市数据中心具有较多优势。
The data center industry has recently seen a demonstrable rise in construction, especially in urban environments near population centers. For data center owners and developers, there are advantages to such proximity.
然而，将数据中心选址定位在城市中，会给项目设计团队带来一些挑战，这些挑战在农村或偏远地区数据中心中不会面临。
Yet positioning data centers in urban environments creates challenges for a project’s design team that it wouldn’t encounter in rural or remote areas.
在过去十年中，云厂商快速增长，建造了大量位于农村偏远地区的数据中心园区。这些云数据中心使用户获得托管虚拟桌面、应用程序和数据存储服务，而数据中心位置在哪并不重要。以前的这种模式是基于所有算力在云中都能够完全商品化的假定。
Cloud provider growth over the past decade has created enormous data center campuses, often in rural areas of the country. These cloud data centers enable users to host virtual desktops, applications and data storage. The initial premise was that it doesn’t matter where these platforms reside. Previously, this paradigm presumed that all computing would be completely commoditized in the cloud.
然而，城市地区的时间敏感型应用程序会产生延迟问题。显然，大多数终端用户和“物联网”设备都位于城市地区，这些设备的数量预计将在近期呈指数增长，因此需要更靠近城市人口中心的数据中心。考虑到美国大部分人口居住在城市地区，这也是合乎逻辑的。但是，当开发商和业主将这些数据中心靠近城市环境时，它们会给设计团队带来一些重大障碍。
However, time-critical applications in urban areas have led to latency concerns. It has become apparent that the majority of end-use and “internet of things” devices are in urban areas. Furthermore, the number of these devices is expected to grow exponentially in the near term, thus creating a need for data centers closer to urban population centers. This is logical, considering that urban areas are where most of the United States population resides. But when developers and owners situate these data centers close to urban environments, they create several significant obstacles for design teams.
数据中心的四大挑战
Four data center design challenges
当决定在市区建立数据中心时，尽早了解四个选址限制对于项目的成功至关重要。
When deciding on a data center site in an urban area, an early understanding of four site constraints is essential to a project’s success.
市政服务
Municipal services
无论数据中心位于何处，园区和建筑物与给水、雨水和排污系统的连接都是基本要求。为满足数据中心选址需求，地方市政系统也有不小压力。在许多情况下，业主和开发商选择工业走廊和科技园区作为城市数据中心的目标位置，因为这些位置更靠近电力和光纤路径。
Campus and building connections to water, storm and sanitary municipal systems are basic requirements regardless of where a data center is located. Proximity to urban environments places pressure on these local municipal systems to meet the projected demands of the site. In many cases, owners and developers choose industrial corridors and technology parks as target locations for urban data centers due to adjacencies to power and fiber paths.
当项目团队对潜在用地进行深入调研时，他们经常发现需要升级市政系统容量。这些市政升级可能需要在场外进行，或作为大型城市项目的一部分，以提升城市供水、雨水或污水设施服务能力来满足预期和未来的需求。
When project teams perform due diligence on a potential property, they often identify a need to upgrade the capacity of a municipal system serving the site. These municipal upgrades may need to happen off-site or be part of a larger city project to increase capacities for water, storm or sanitary services for expected and future demand.
此外，目前水比电便宜，这推动了多个地区的超大规模数据中心使用蒸发冷却系统。高用水量也可能引起社区担忧，尤其是当拟建的数据中心在缺水地区每天可能使用数百万加仑的水时。
In addition, water is currently cheaper than electricity and this has driven up the use of evaporative cooling systems for hyperscale data centers in several regions. High water usage may generate community concerns as well, especially when a proposed data center may use millions of gallons of water a day in a water-stressed region.
空气许可
Air permitting
另一个考虑因素是环保局的“环境正义”计划。数据中心通常会配备大量柴油发电机，以在电力设施中断时为数据中心提供支持。因此，只有获得了环境空气许可，数据中心才能运行这些发电机。
Another consideration is the Environmental Protection Agency Environmental Justice program. A data center will generally have a significant number of diesel engine electrical generators to support the data center in the event of an electrical utility outage. As such, these engines will require environmental air permitting for the data center to operate the engines.
运营许可证是许可机构向备用发电机等空气污染源设备发放的具有法律效力的文件。在城市地区，如果数据中心位于低收入和/或少数民族人口的一定距离内，它将成为一个令人关注的“环境正义”领域。这意味着会通知居民，可能会举行公开会议，这可能会给数据中心的许可过程带来技术和公关双重复杂性。
Operating permits are legally enforceable documents that permitting agencies issue to air pollution sources such as standby engine generators. In urban areas, if the data center is located within a certain distance of low-income and/or minority populations, it becomes an Environmental Justice area of concern. This means residents are notified, there may be public meetings and this could introduce technical and public relations complexities to the permitting process for the site.
湿地
Wetlands
位于数据中心选址中的湿地区域始终是不理想的。潜在的城市数据中心选址中，确实会存在较小的湿地区域，它们带来了与农村地区类似的挑战。根据数据中心园区设计和总体规划的土地需求，业主可能需要购买湿地积分来抵消移除或部分移除现有湿地的影响。
Wetland areas situated on a data center site are never ideal. At potential urban data center sites, smaller wetland areas do exist and they pose similar challenges like those found in rural areas. An owner may need to purchase wetland credits to offset impacts, removal or partial removal of the existing wetlands on-site based on the data center campus design and masterplan land requirements.
湿地缓解信用是一种贸易单位，用于抵消美国水域发生的生态损失，由美国陆军工程兵团和EPA监管。另一个挑战是现场施工物流，尤其是在现有湿地周围进行施工的情况下。
A wetland mitigation credit is a unit of trade used to offset ecological losses that occur in waters of the U.S., which are regulated by the U.S. Army Corps of Engineers and EPA. Another challenge is site construction logistics, especially if construction takes place around existing wetlands.
此外，通过陆军工程兵团或EPA审查湿地影响的时间有时是不可预测的，在项目团队等待许可批准期间可能会延长项目时间表。
In addition, overall permitting timelines to review the impact to wetlands through the Army Corps of Engineers or EPA are sometimes unpredictable and may elongate project schedules while the project team is awaiting permit approvals.
电力设施
Electrical utilities
数据中心是巨大的电力消耗者。目前估计，全球1%的电力用于云计算，预计这一比例将在未来十年呈指数级增长。而人工智能、增强现实和数字货币的出现，可能会继续推进这种电力消耗。
Data centers are enormous consumers of electricity. It’s currently estimated that 1% of the world’s electricity goes to cloud computing and this percentage is expected to grow exponentially over the next decade. The future demands of artificial intelligence, augmented reality and the advent of digital currencies will likely continue to drive this power consumption.
最初，在城市中心附近处理更多的数据是有意义的，因为这是需求所在。然而，当计算需求增加时，计算效率会降低，数据中心可能无法完全满足预期的云计算需求。
Initially, it will make sense to process more data near urban centers, as this is where the demand will be. However, it may not be possible to support the sheer magnitude of the projected cloud power requirements, as increased computing needs couple with decreasing efficiencies.
电力供应一直是数据中心选址的诸多驱动因素之一，且在短期内可能会成为主要因素，因为电力供应很难通过新建发电厂和基础设施迅速满足预期的需求增长。
Available utility power has always been one of many drivers for data center locations, but it may end up being the main one in the near term because it will be challenging for utilities to quickly meet the anticipated rise in demand via new power plants and infrastructure.
电力公司通常需要相当长的时间来敲定服务协议，协商负载斜坡可用性的时间，延长电力服务线或建造新的变电站。在城市地区，通常会有来自其他数据中心或工业邻居的类似资源竞争。例如，如果在数据中心周围规划一个大型物流站点，为电动汽车充电的卡车车队可能会与数据中心竞争电力可用性。
Electrical utilities often take a considerable amount of time to finalize service agreements and negotiate the timing of load ramp availability, extend electrical service lines or build new electrical substations. In urban areas, there is often competition for the similar resources, either from other data centers or industrial neighbors. For example, truck-fleet charging for electric vehicles could be competing for the electric utility availability if a major logistics site is planned for the neighborhood surrounding the data center.
交付速度一直是数据中心的重要驱动力。然而，有时项目受制于场地公用电源的可用性和时间安排。随着电力需求的增加，人们对通过太阳能光伏板和风力涡轮机等绿色替代品现场发电技术产生了新的兴趣。现场光伏电池板和风力涡轮机大规模运行占地面积大，而数据中心通常面临城市环境中可用空间的挑战。
Speed-to-market has always been an important driver for data centers. However sometimes a project is beholden to the availability and timing of utility power to a site. As electrical demand increases, there is a renewed interest in on-site generation through green alternatives such as solar photovoltaic panels and wind-powered turbines. On-site PV panels and wind turbines need a lot of real estate to function at scale and data centers are usually challenged with available space in urban environments.
另一种现场发电方案是使用最新的燃气轮机技术。用于数据中心现场发电的燃气轮机正在成为一个有吸引力的选择，因为它们将解决当前以及可能长期持续的公用电网挑战。未来，预计燃气轮机将完全使用氢等无碳燃料。因此，在已存在电网挑战的城市地区，许多开发商将会探索如何将现场发电纳入未来数据中心的规划中。
Another on-site generation option is the use of the latest gas turbine technology. Gas turbines for generating power on-site for data centers is becoming an attractive option as they would solve current utility grid challenges and those that may continue in the long term. In the future, gas turbines are predicted to run exclusively on carbon-free fuels such as hydrogen. It is likely, therefore that many developers will explore ways to incorporate this type of on-site generation into plans for future data centers in urban areas with known grid challenges.
筹备新的数据中心
Preparing for a new data center
在城市地区考虑潜在的数据中心选址时，及早了解关键场地限制对于项目的成功至关重要。需要分析以下所有因素：确定有关市政给水、雨水、污水基础设施系统连接可用性的限制因素；了解场地的空气许可或湿地环境限制；以及与服务电力公司的协调。
When selecting a potential data center site in an urban area, an early understanding of significant site constraints is essential to a project’s success. The following all factor into this analysis: identification of limiting factors regarding the availability of the municipal infrastructure systems, water, storm and sanitary connections; understanding a site’s air permitting or wetland environmental constraints; and coordination with the serving electrical utility.
项目团队通常需要花费大量时间和调研发现选址限制，减少选址对城市数据中心项目成本或进度的潜在影响。
It often takes a considerable amount of time and research by the project team to uncover site constraints and mitigate any potential impacts to an urban data center’s cost or project schedule.
深 知 社
翻译：
任文静
哔哩哔哩 基础工程部项目管理
DKV（Deep Knowledge Volunteer）成员
贾梦檩
阿里云 暖通工程师
DKV（Deep Knowledge Volunteer）成员
|