答疑解惑 | 电力能被软件定义吗？
Can power be software-defined?
关键的基础设施通常是不灵活的。安迪·劳伦斯认为数据中心的能源真的可以作为一种灵活的服务来使用吗？
Criticalinfrastructure is normally inflexible. Andy Lawrence finds out if data centerpower really can be delivered as a flexible service
January 31, 2019 By Andy Lawrence, UptimeInstitute Comment
2011年左右的时候，一个新的词进入了IT领域: '软件定义的.' 这个术语有时被认为出自于一个VMware的工程师，通常出现在'软件定义的数据中心'或者'软件定义的网络'的语境里。 这个奇怪的措辞描述如何使用分布式的管理架构，提出介绍了一种新型的弹性水平对于一个网络或一池子的计算资源是可能的，重建和撤回连接，或重新部署计算资源，并且可以即时生效随时更新。
Sometime around 2011, a new phrase entered the IT lexicon:'software-defined.' The term is sometimes credited to an engineer at VMware, and usuallyappears in the context of the 'software-defined data center' or'software-defined network'. The strange phrasing describes how, usingdistributed management architecture, it is possible to introduce a new level of'plasticity' to a network or a pool of compute resources, re-making andun-making connections, or re-allocating compute resources, on the fly.
维基百科上描述软件定义的数据中心是一个市场营销术语，它或许反映了这样一个现实，在软件定义的商品贸易列表中的公司。 事实上他们可能被认为是创新的,云导向的,有颠覆性的，因此十分有趣。 许多早期的软件定义的创新者在非常高的估值被收购了。
Wikipedia describes SDDC as a 'marketing term,' which perhapsreflects the reality that companies and products in the software-definedcategory trade on the fact that they might be considered innovative,cloud-oriented, disruptive and therefore of great interest. Many of the early'software-defined' innovators were acquired at sky-high valuations.
在这样的语境下，被放在软件定义上的价值和重要性没有不合时宜的。这个架构考虑到许多IT元器件层的硬连线的复杂性被抽象和转移到了一个管理的层面，因此考虑到更简单,同质/同样的设备被定期的部署和管理。聚合资源的使用对云计算的经济性是关键的。
The value andimportance put on 'software-defined' in this context was not misplaced. Thearchitecture allows for much of the hard-wired complexity at the componentlevel in IT to be abstracted out and moved to a management 'plane,' therebyallowing for simpler, homogeneous devices to be routinely deployed and managed. This use of aggregatedresources is critical to the economics of cloud computing.
但同样重要的是，软件定义的架构使得在大型及更大的数据中心中能更加容易和便宜的管理高度复杂和动态的能量潮流变成可能。没有软件定义的网络的技术，5G网络和物联网边缘网络将很可能几乎无法(不可能)建设而且也不经济。
But equallyimportant, software-defined architectures make it possible to much more easilyand cheaply manage the highly complex and dynamic 'east-west' traffic flows in largedata centers and beyond. 5G networks and IoT edge networks would likely be nearimpossible to build, and uneconomic, without software-defined networkingtechnologies.
有时，一些数据中心的创新者一直致力于如何把软件定义的模型被应用在点电能上。这个在IT上具有颠覆性和革命性的改变能否在电能的分发和管理方式上有类似的效果？ 毕竟电能相比于比特流而言没有很大的不同：它是一个电子流能被存储，生成和消耗，也能被分布于一个网络，是可以切换（用开关控制）和路由的。因此是可以从远程管理的。 对此的一个后续的问题是：即使电能能被以这样的方式管理，(就算)考虑到包含许多特别复杂和额外支出的电力管理的情况上。什么时候和情况下这样做是安全的，经济的和有益处的？
For some time, someinnovators in the data center have been working on how the model of“software-defined” can be applied to power. Does this disruptive andrevolutionary change in IT have an equivalent in the way that power isdistributed and managed? Power is, after all, not so dissimilar to the flow ofbits: it is a flow of electrons, which can be stored, generated and consumed,is distributed over a network, and is switchable and routable, and thereforemanageable, from a distance. A follow-on question to this is: Even if power canbe managed in this way, when is it safe, economic and useful to do so, giventhe many extra complexities and expenses involved in managing power?
以上两个问题的回答是电能和在IT层面上的二进制和确定性是不同，在IT层面上软件定义的影响是及时的，巨大的和完全颠覆的。而在更多智能和自动化应用的时候在几乎所有市场上是明显有用的。在能源分发上的创新机会相比IT而言是更不明确的。 数据中心是一个很突出的案例。 对于供应商，大型设施出现了一个很大的机会，由于市场的规模和成长。 大量预留空间，高花费和不灵活性在电能的分发和使用上。但是与此同时，运营者对于高效和敏捷的渴望可能被客户的惯性，现存的投资的资产，僵化的设计和商业模式强烈的限制了，对于高可用性解决方案的需求已经被证明是过时的了（低风险的考量）。
The answer to bothquestions is not quite as binary or as definitive as at the IT level, where theimpact of software-defined has been immediate, significant and clearlydisruptive. While the application of more intelligence and automation isclearly useful in almost all markets, the opportunities for innovation in powerdistribution are much less clear. Data centers are a stand-out example: Largefacilities present a great opportunity for suppliers, because of the size andgrowth of the market, vast over-provisioning, high costs, and inflexibility inpower distribution and use. But at the same time, the operator’s aspirationsfor efficiency and agility may be strongly constrained by customer inertia,existing investments, rigid designs and business models, and the need for highavailability solutions that have been proven over time (low risk).
在更大的能源市场，然而，许多平行和连续的创新浪潮已经席卷了行业，因为该行业走向更加动态和分布式的发电和存储，更好的用上了智能化和自动化以及更扁平和交互的模式。供应商进入了能源管理和存储 – 举几个来说 ABB, GE, Eaton, Siemens, Schneider Electric and Vertiv, 他们已经正在开发智慧能源技术，在很多情况下会持续数十年。 新加入者 – 最突出的是电动车和电池创新者特斯拉- 也已经提出了激进的新的储能和配电技术。这个影响正在公用电网方面被看到， 然而微电网，现在的一个成熟的技术但一个年轻的市场，提供了一个动态和智能的电源，存储和消耗的管理的模型。
In the wider energymarket, however, a number of parallel and successive waves of innovation havebeen sweeping through the industry, as the sector moves to more dynamic anddistributed generation and storage, greater use of intelligence and automation,and flatter, transactive models. Suppliers working in the field of powermanagement and storage - ABB, GE, Eaton, Siemens, Schneider Electric andVertiv, to name a few - have been developing “smart energy” technologies, insome cases, for decades. New entrants – most notably electric car and batteryinnovator Tesla – have also introduced radical new storage and distributiontechnologies.
The impact is beingseen at the utility grid level, while microgrids, now a mature technology but ayoung market, provide a model for dynamic and intelligent management of powersources, storage and consumption.
在数据中心，相似的创新正在进行中，虽然采用情况是零散局部的。范围从先进能源管理系统的引进（这可能被用在监控能源的使用和告知关于在采购，存储和消耗的决策上）到微电网技术。 电源和IT的消耗处于中央控制下被监控和管理（目前，这项技术在一个数据中心的环境下很少采用以及最可能在一个有高性能计算中心的校园中被看到）
In the data center, similar innovations are underway, althoughadoption is patchy. These range from the introduction of advanced energymanagement systems, which may be used to monitor energy use and informdecisions about purchasing, storage and consumption, to microgrid technologies,where power sources and IT consumption are monitored and managed under centralcontrol (at present, this technology in a data center context is rare and is mostlikely to be seen on a campus with a high performance computing center).
但是最明显的与软件定义的数据中心的关联/比较是，软件定义的能源SDP架构正由一个硅谷小公司Virtual Power System (VPS)推广； 相比更不明显的是，先进的共享储备架构的使用，正在被一些UPS供应商们推广，最突出的是Vertiv. 这些架构是非常不同的，在两种情况下，其中最关键的目标之一是减少多余的容量以及转移或存储能源给到各种最需要用的应用、机架上，同时避开那些需要最少的设备。
But the most obvious connection/comparison with thesoftware-defined data center is the software-defined power (SDP) architecture, aspromoted by the small Silicon Valley company Virtual Power System(VPS); and, less obviously, the use of advanced, shared reserve architectures,as promoted by some UPS suppliers, most notably Vertiv. These architectures arevery different, but in both cases, one of the key goals is to reduce the needfor spare capacity, and to divert or reserve power for applications, racks or rows that mostneed it, and away from those that need it least.
VPS的架构是很特别的：它利用了分布式的锂离子电池在机架上，提供了一个虚拟的分布式的能被在需要的时候使用和管理的能源池上，或提供作为备用。 在这个意义上，它和在软件定义的数据中心中的同质计算资源池是类似的。 VPS部署很多小型的中心化管理的，安装在机架上的控制设备。 俗称ICE 开关，它能被用来关闭在机架上的主要的UPS电力从而利用上本地的锂离子电池。
VPS’ architecture is quite specific: it makes use of distributedlithium-ion batteries in the rack, to provide a “virtual” distributed pool ofpower that can be used and managed when needed, or to provide back up. In thissense, it is analogous to the homogeneous pool of compute resources in a software-defined data center. VPSdeploys a number of small, centrally managed, rack-mounted control devices,known as ICE switches, which can be used to turn off the main UPS power at therack and thereby draw on the local Li-ion battery.
管理软件扮演了一个重要的角色。 它能有效地转移电力从一个机架到另一个。 不是通过使用锂离子电池从一个机架去给其他地方的机架供电（虽然这是可能的，但这会更加复杂因为电力转换和协调的要求）但是通过从中央UPS到某一机架上的电池的转换，使得更多在其他地方的电力变得可得到。
The management software plays a key role. Effectively, it candivert power from one rack to another – not by using the Li-ion batteries fromone rack to power another rack elsewhere (although this is possible, it is morecomplex because of the power conversion and harmonization requirements), but byswitching from the central UPS to battery in certain racks, making more poweravailable elsewhere.
为了作这个决定，管理软件根据服务器负载的性质和保护要求使用不断变化的数据。 最后虽然在早期的几天，负载可能会被转移去匹配能源的可用性或转移为了减少电力在其他地方的使用。就像在一个软件定义的网络中，中央软件正在使用数据和策略去智能控制相对简单的设备.
In order to make such a decision, the management software usesever-changing data about the nature of the loads on the servers and the levels ofprotection required. Ultimately, although in its early days, loads may be movedaround to match the power availability, or moved in order to release power foruse elsewhere. As in a software-defined network, the central software is usingdata and policies to intelligently control relatively simple equipment.
在一个软件定义的能源环境下，软件可能被认为是一个通往动态资源充足的数据中心的步骤， 其能力会被保留给最关键的应用，而其他相对不重要的应用可能会有更少容量被分配 （“dynamic tiering” 不是一个恰当的术语，因为这是关于电力可用性不是容错性或可维护性的意思）
In an SDP environment, the software might be considered as astep on the road to “dynamically resourced” data centers, with capacity reservedfor the most critical applications, while other less-important applications mayhave less capacity allocated (“dynamic tiering” is not the appropriate term, asthis is about power availability, not fault-tolerance or maintainability).
软件定义的电力软件也可以用电池去提供额外的充足的能源在用电高峰时期，能有效的使得机房去使用更多的电力比起不间断电源的供给。或者软件也能被用来去使运营者使用本地电力当电网处于最贵的时期成为可能，或者当主机托管的客户希望去避免超过协议用电限制,触发额外的收费的时候
SDP software can also use the batteries to offer an extra,supplemental power source during times of peak demand, effectively enabling adata hall to use more power than has been provisioned by the UPS; or it couldbe used to enable operators to use local power at times when the grid power ismost expensive, or when colocation customers wish to avoid going beyond agreedpower use limits, triggering extra charges.
Uptime Institute 认为，VPS体系架构只是数据中心里可能被归类为“智能电力”或“智能能源”的许多方法中的一种，尽管这些用例并不完全相同。例如，集中式UPS系统采用“三比二”配置，甚至多达“八比七”配置，可以使用智能的、策略驱动的软件转换来维持容量，分散负荷，降低风险。
Uptime Institute sees the VPS architecture as only one of anumber of approaches in the data center that might be classed as “Smart Power”or “Smart Energy,” although not at all the use cases are the same. For example,centralized UPS systems that pool capacity in a “three makes two” configuration, or even up to“eight makes seven” configuration, can use intelligent, policy-drivensoftware switching to maintain capacity, spread loads and reduce risk.
行业的趋势是朝着减少预置的成本，使用N 1或 N x的配置（大约所有数据中心的40%有N 1的解决方案）,而不是完整的2N配置。 但是这带来一个风险：Uptime Institute Research’s 2018 行业调查发现2N架构的数据中心经历了更少的断供比起那些N 1架构的。 在2N架构的运营者里，35%经历过一次断供在过去三年，而51%N 1架构的运营者同期有一次断供。
The industry trend is towards reducing the provisioningcosts using N 1 or N x configurations (about 40 percent of all datacenters have an N 1 solution), rather than full 2N configurations. But thiscarries a risk: Uptime Institute Research’s 2018 industry survey found thatdata centers with 2N architectures experience fewer outages than those with N 1architectures. Among the operators with a 2N architecture, 35 percentexperienced an outage in the past three years, while 51 percent of those withan N 1 architecture had an outage in the same period.
这个的可能是这不是完整的由于在多余容量上的投资。这也许有争议因为额外的设备被增加用来补足更多的颗粒解决方案，额外的连接和交换机被用来去连接额外的UPS系统，好的管理，容量的计划和维护的重要性变得更加重要了。
The likelihood is that this is not entirely due to theinvestment in the extra capacity. It may be argued that as extra equipment isadded to make up more granular solutions, and extra connections and switches are usedto link up the extra UPS systems, the importance of good management, capacityplanning and maintenance becomes more important.
这是在软硬件结合诸如共享保留的系统或软件定义的系统可能进入到他们本身的地方。共享保留的系统可以将很多更小的UPS系统的容量集中起来，然后通过软件管理的转换开关将电源从共享备用系统中动态分配给PDU。 这牵涉到一些复杂的转换，自然地会让想要确保简单的运营和管理的管理者担心。但关键是软件-如果这是策略驱动的，易于操作的，有恢复/复原的设计，那么可靠性应该是非常高的-同时当随着专业知识和AI技术不可避免的嵌入的时候变得更高。
This is where hardware/software combinations such as sharedreserve systems or software-defined systems may come into their own. Sharedreserve systems can be used to pool the capacity of multiple smaller UPS systems, and thendynamically allocate power from this pool to PDUs using software-managed transferswitches. This involves some complex switching, andnaturally worries managers who want to ensure ease of operation and management.But the key is the software – if it is policy-driven, easily operated, anddesigned with resilience, reliability should be very high - and becomes higherover time as expertise and AI techniques are inevitably embedded.
如果可以在不增加风险或成本的情况下部署电源管理和控制软件（和相关的设备），并且可以在很大程度上与直接使用硬连接线物理设备所能达到的可用性水平相匹配，那么软件定义的能源的方法的论证是清楚的：更大的灵活性在能源如何被引导，管理和部署上。反过来。相对于工作量，能源容量可能会大大增加。
If power management and control software (and associatedequipment) can be deployed without adding significantly to either risks orcosts, and can largely match the proven levels of availability that can beachieved by directly using hard-wired physical equipment, then theargument forsoftware-defined power approaches is clear: greater flexibility in how power isdirected, managed, and deployed. In turn, power capacity, relative to theworkload, may be significantly increased.
最终，这可能意味着数据中心管理者，正在使用一系列工具（应用），拥有中心化的控制在电力的如何被使用，分发，限制，存储或甚至被出售来满足一些变化的需求上，服务等级或者策略上。它应该会使得更加精益的各种大小的数据中心的建造成为可能。供应商诸如Schneider和Vertiv在其中涉及到开发工作，和更小的供应商比如VPS的合作。 同时主要的运营商包括Equinix正在调查技术的价值。同时，更小的中心化管理的UPS系统集群和其他电力设备正在被使用去创造更多颗粒状的和可管理的能源存储。随着时间流失，更多普遍的采纳这些技术，不管供应商是谁以及不管技术怎么被称呼，看起来都是很有可能的。
Ultimately, this could mean that the data center managers, usinga suite of tools (applications), have some centralized control over how poweris used, distributed, capped, stored, or even sold to meet changing demands, service levels,or policies. It should enable the building of leaner data centers of all sizes.Suppliers such as Schneider and Vertiv are among those involved in thedevelopment effort, working with smaller vendors such as VPS; meanwhile majoroperators, including Equinix, are investigating the value of the technology. Atthe same time, smarter, centrally managed constellations of UPS systems andother power equipment are being used to create more granular and manageablereserves of power. Over time, more widespread adoption of these technologies,whoever the suppliers are and whatever the technology is called, seems very likely.
但在采用上存在障碍。事实上，在Uptime Institute和451 research的“颠覆性数据中心”研究项目中，一组专家给SDP的排名最低（3.4分\满分5分），包括可能的使用率和受影响程度，以及对行业的影响。 更大批的参与调查的数据中心运营者甚至没有什么热情。
But there are barriers to adoption. In fact, in the “DisruptiveData Center” research project by Uptime Institute and 451 Research, a panel ofexperts gave SDP one of the lowest rankings (3.4 out of 5) for the likelyrate and extent of take up, and the likely impact on the sector. A largerpool of data center operators that we polled was even less enthusiastic.
这个怀疑最可能反应了以下原因的结合包括：不熟悉，和可能在论证这样的投资的困难性，尤其是现存的环境是已经很稳定的以及这里的许多花费是在贬值的。 但在早期怀疑没有被消除和大范围的采纳：直接液冷和微电网的评分甚至都低于软件定义的能源，然而有一个案例在论证这两个技术都有一个很好和长期的未来。对于这些技术的布道者，部分的挑战在于说服运营者在一个如此十分落地的现存架构体系的商业环境下去投资。
This skepticism most likely reflects a combination ofunfamiliarity, and the possible difficulty of justifying such investments,especially for existing environments that are already stable and where many ofthe costs have been depreciated. But this early stage skepticism does not ruleout later, widespread adoption: direct liquid cooling and microgrids bothscored even lower than SDP, yet there is a case for arguing both have a strong,long-term future. For evangelists of these technologies, part of the challengeis in convincing operators to invest in a business environment so heavilygrounded in existing architectures.
SDP可能最好被视为一套独立但相互关联技术的一部分，在为其它技术铺平道路之前，每个技术都需要取得进展。其中一个重要的是采用锂离子的（或其他类似的）电池化学物质，无论在中心还是在机架上；比起铅酸电池，这些电池能被放电和再被充电成百上千次，而对电池容量或寿命没有太大影响。这为更动态、更智能地利用能源开辟了道路。
SDP is probably best viewed as part of a package of separate butinterlinked technologies that each need to make progress before paving the wayfor others. A key one of these is Li-ion (or similar) battery chemistryadoption, both centrally and in the racks; these batteries can be dischargedand recharged hundreds to thousands of times without much effect on batterycapacity or life compared to lead-acid batteries, opening the way for much moredynamic, smarter use of energy.
相同的是，诸如DCIM,软件自动化和AI技术被许多运营者谨慎的看待。作为更智慧，软件驱动的工具和系统变得更加成熟，智能和集成。远程控制的使用，自动化程度的增加，更敏捷的软件驱动的系统的采纳都将会增加。这些系统有望带来更加高效和更好的容量的使用，减少维保并且增加容错性。
Equally, technologies such as DCIM, software automation and AIare regarded cautiously by many operators. As smarter, software-driven toolsand systems become more mature, more intelligent, and more integrated, and theuse of remote control and automation increases, the adoption of more agile,software-driven systems will increase. Such systems promise greater efficiencyand better use of capacity, reduced maintenance, and increased fault tolerance.
翻译：
Leaf Zhu
General Manager of ICT-Event
校对：
邓秋实
中科大 国家同步辐射实验室
DKV（Deep Knowledge Volunteer）创始成员