百度地图开放平台计算距离_开放计算项目可能会对未来的数据中心做出贡献

百度地图开放平台计算距离

As the leader of the Open Compute Project, Facebook has opened up many of the designs that it uses for servers and equipment in its data centres for use by anyone. The intended purpose is to create “one of the most efficient computing infrastructures at the lowest possible cost” and since its inception 2 years ago, we have seen many other hardware manufacturers contribute their hardware designs to the project in an effort to lower the cost of building a highly powered infrastructure by removing the middleman. It is indeed very feasible that the Open Compute Project could be the basis of the data centres of the future given the necessity of many companies to lower their IT costs without compromising on performance.

作为开放计算项目的领导者，Facebook已经开放了其用于数据中心中的服务器和设备的许多设计，供任何人使用。 其预期目的是创建“以尽可能低的成本获得的最高效的计算基础架构之一”，自两年前成立以来，我们已经看到许多其他硬件制造商为该项目贡献了其硬件设计，以降低硬件成本。通过移除中间商来构建功能强大的基础架构。 鉴于许多公司有必要在不影响性能的情况下降低IT成本，因此将开放计算项目作为未来数据中心的基础确实是非常可行的。

Energy efficient devices

节能设备

The Open Compute Project was founded by Facebook as a way of sharing the custom designs that it uses for hardware in its data centres, with the sole intention of creating a more efficient infrastructure that is able to collectively deliver more power at a lower cost on the same footprint. It is estimated that utilising these designs rather than purchasing hardware direct from OEMs could enable organisations to realise increased efficiencies of up to 38% and whatever the size of the deployment, this will remove a significant and noticeable chunk of the existing costs. Whilst currently the main focus has been put on server designs, contributors to the Open Compute Project, which include IBM and Microsoft to name but a few, are looking to develop more efficient designs for other data centre devices – the current target of which is network switches.

开放计算项目由Facebook创立，旨在共享其用于数据中心硬件的定制设计，其唯一目的是创建一个更高效的基础架构，从而能够以较低的成本集体提供更多的电力。相同的足迹。 据估计，利用这些设计而不是直接从OEM购买硬件可以使组织实现高达38％的效率提高，并且无论部署规模如何，这都将节省大量的现有成本。 虽然目前的主要重点是服务器设计，但Open Compute Project的贡献者(包括IBM和Microsoft等)正在寻求为其他数据中心设备开发更高效的设计-当前的目标是网络开关。

Cooler servers

散热服务器

For the most part, the increased efficiencies arise from the use of better cooling designs. Facebook’s own designs have used a 1.5U design, which is enough to fit in taller heat sinks and larger fans so that heat can be dissipated much faster in a more efficient manner. Anandtech notes that these improvements mean that the fans only use between 2% and 4% of the total power consumed by the server, an incredible figure when you consider that the fans in a standard 1U server can use up to 33% of the unit’s total power. This makes the 1.5U design just as efficient as the most efficient 3U servers that we see on the market.

在大多数情况下，效率的提高是由于采用了更好的冷却设计。 Facebook自己的设计使用的是1.5U设计，足以容纳更高的散热器和更大的风扇，从而可以更有效地更快地散热。 Anandtech指出，这些改进意味着风扇仅消耗服务器总功耗的2％至4％，当您考虑到标准1U服务器中的风扇最多可使用设备总功耗的33％时，这是一个令人难以置信的数字功率。 这使得1.5U设计的效率与我们在市场上看到的最高效的3U服务器一样高效。

A transferrable rack design

可转移的机架设计

Another core component of the Open Compute Project is what is dubbed the ‘open rack’, a universal rack design that is to consider other elements that are necessary for the running of data centre hardware such as power distribution, battery backup, cable organisation and thermal design/server cooling. The intention of this rack design is for it to be shared across a number of data centres and that servers produced using the Open Compute designs should be able to just slot into the rails on the rack without the need for other accessories, as is the case at the moment with many OEM server designs. Overall this should let network technicians transport servers and other Open Compute-designed hardware between racks in a fraction of the time it takes to transport current hardware models.

开放计算项目的另一个核心组件是所谓的“开放机架”，这是一种通用机架设计，旨在考虑数据中心硬件运行所必需的其他要素，例如配电，电池备用，电缆组织和散热设计/服务器散热。 这种机架设计的目的是要在多个数据中心之间共享，并且使用Open Compute设计生产的服务器应该能够仅插入机架的导轨中，而无需其他附件。目前具有许多OEM服务器设计。 总体而言，这应该使网络技术人员可以在机架之间传输服务器和其他Open Compute设计的硬件，而所需的传输时间仅为传输当前硬件模型的时间的一小部分。

Cheaper servers (for larger enterprises at least)

便宜的服务器(至少适用于大型企业)

By cutting out the middleman in the hardware purchase process, many large organisations that choose to adopt Open Compute designs and create their own servers could stand to save a lot of money. Not only will they be able to save on the amount being spent on rack space since Open Compute Project designs allow for more physical servers in the same amount of space, but you will be removing the premium charged for “pretty” servers by OEMs because you will be purchasing the individual server components yourself. Then there is of course the freedom to create servers with varying specifications to match the intended purpose – the best thing here is that you will not be restricted by the limitations imposed by some OEMs.

通过在硬件购买过程中减少中间商，许多选择采用Open Compute设计并创建自己的服务器的大型组织可以节省很多钱。 由于Open Compute Project设计可以在相同的空间量内容纳更多的物理服务器，因此它们不仅可以节省机架空间上的花费，而且您将省去OEM为“漂亮”服务器收取的额外费用，因为您将自己购买单个服务器组件。 然后，当然可以自由地创建具有不同规格的服务器以符合预期的目的–最好的是，您将不会受到某些OEM施加的限制的限制。

My opinion is that cheaper servers can only help to improve the pricing of cloud computing and as such it shall become more accessible to smaller businesses, especially those from developing countries. In turn this means that businesses of all sizes have access to powerful computing platforms, giving them the opportunity to prosper in the digital age.

我的观点是，便宜的服务器只能帮助改善云计算的价格，因此，小型企业(尤其是来自发展中国家的企业)将更容易使用它。 反过来，这意味着各种规模的企业都可以使用强大的计算平台，从而为他们提供了在数字时代繁荣发展的机会。

翻译自: https://www.eukhost.com/blog/webhosting/contribution-open-compute-project-data-centres-future/

百度地图开放平台计算距离