人们出于各种目的使用计算机。你家里的电脑可以做很多事情。你可以浏览网页、玩视频游戏和(可悲地)完成一些工作。我们大多数人所做的工作归结为键入文档或摆弄一些Excel电子表格。你甚至可以做一些内容创作,比如视频编辑。
普通的现代家用电脑可以在一定程度上完成所有这些事情。然后是专用的工作站计算机。专业级设备,用于从专业录音到复杂的科学和工程模拟的任何事情。
在高端,这些桌面工作站可以运行到数万美元。当然,他们最终会通过他们所做的工作为自己付出代价,但这可能需要一段时间——而且前期成本很高。
但现在可能有另一种方式来获得强大的计算能力——云计算。
什么是基于云的模拟?
我们大多数人每天都在使用基于云的计算服务。如果您使用Gmail、DropBox或Office 365,您正在访问远程计算机资源。这种形式的云计算称为SaaS或软件即服务。(Software as a Service. )
这是提供云计算的一种方式。您还可以在组成云数据中心的大量功能强大的计算机上租用时间。然后运行您喜欢的任何软件。
基于云的模拟可作为SaaS产品使用。将这些视为您可能在功能强大的工程或科学工作站上运行的软件的云版本。您可以使用普通计算机或移动设备访问该软件的界面。
您对模拟进行编程或上传数据,然后数百个处理器会在短时间内开始运行以处理数字。几秒钟或几分钟后,您的模拟结果就可用了。
基于云的模拟示例
基于云的模拟服务正在兴起,似乎每隔几个月我们就会听到一项新服务。虽然哪种服务可能适合您不在本文的讨论范围内,但有一些主要示例可以很好地展示这种新技术应用程序的多样性和广度。
上图所示的SimScale(SimScale)供工程师、科学家和设计师使用。从虚拟风洞到流体动力学和机械应力测试。您可以通过平板电脑完成所有操作,而他们的服务器场则负责所有繁重的工作。
AutoDesk提供其行业标准CAD 软件(CAD software)的基于云的版本。CAD工作站是专业计算机的主要类别,但使用云解决方案您不必担心,例如拥有昂贵的工作站 GPU(workstation GPU)。
Ansys和Rescale都通过云提供HPC(高性能计算)服务。这就像租用自己的超级计算机,只要您需要它来完成特定的工作。
这只是几乎任何人都可以使用的令人难以置信的计算能力的冰山一角。
本地工作站的优点
使用基于云的模拟解决方案而不是使用您自己的本地专业计算机对每个人来说都不是最好的主意。事实上,现在最好只让少数关键类型的专业用户跳到云端。
那么为什么要坚持使用本地工作站呢?这种方法的第一个重要优势是数据隐私。如果您使用任何类型的第三方计算工具,您的数据就会失控。因此,如果正在对非常敏感的事情进行工作,您必须 100% 确保服务提供商的安全措施和隐私政策符合要求。
本地工作站选项的另一大优势是您可以在其上运行大量软件。几十年来,工作站一直是主要的解决方案。所以大部分的支持仍然存在。事实是,您的小众专业软件包可能尚未在云端。
云模拟的优点
当谈到走云模拟路线时,这种解决热计算机硬件难题的新方法有很多话要说。
一方面,基于云的模拟没有很大的前期成本。您只需为您使用的内容或包含一定使用量的固定订阅付费。无论哪种方式,都可以将成本降到最低。把它想象成电。您拨动开关并为您消耗的电量付费。
当您使用本地计算能力时,您必须支付所有硬件、维护和升级费用,而大部分时间不是 100% 使用机器,并且可能需要它在关键时刻提供超过 100% 的服务。
流动性(Mobility)是另一个重要因素。几乎(Virtually)任何具有互联网连接和现代网络浏览器的设备都可以使用基于云的服务。这意味着您不必随身携带大型工作站级笔记本电脑或便携式台式机解决方案。工程师可以在现场拿出 iPad 并随意运行模拟,这真是太神奇了。
云并不适合所有人
当然也有缺点。例如,服务的可靠性完全掌握在服务提供商手中,这使得良好的服务水平协议至关重要。对于个人用户,这意味着您必须为不可避免的中断做好计划。即使一项服务有 99% 的正常运行时间,它也会(will)在某个时候出现故障。
你还需要一些相当强大的互联网带宽来移动这些数据,并相信公司会在他们手中保护它。如果这些对您来说似乎不是问题,那么与其升级这些工作站,不如在虚线上签名。
Do Cloud-Based Simulation Services Mean You Can Ditch That Expensive Workstation?
People use computers for all sorts of purpоses. The computer you have at home can do a vаrіety of things. You can browse the web, play video games and (sadly) get some work done. The sort of work most of us do comes down to typing documents or maybe fiddling with a few Excel spreadsheets. You might even do a little content creation, like video editing.
The average modern home computer can do all of these things to one extent or another. Then there are dedicated workstation computers. Professional-grade devices that are used for anything from professional sound recording, to sophisticated scientific and engineering simulations.
At the high end, these desktop workstations can run into the tens of thousands of dollars. Of course, they eventually pay for themselves through the work that they do, but it can take a while – with a substantial upfront cost.
But there now may be an alternative way to access serious computing power – cloud computing.
What Is Cloud-Based Simulation?
Most of us already use cloud-based computing services on a daily basis. If you use Gmail, DropBox or Office 365, you’re accessing remote computer resources. This form of cloud computing is known as SaaS or Software as a Service.
That’s one way cloud computing is offered. You can also rent time on the vast array of powerful computers that make up the cloud data centers. Then run whatever software you please.
Cloud-based simulations are available as SaaS offerings. Think of these as cloud versions of the software you might have run on a powerful engineering or scientific workstation. You can use a normal computer or mobile device to access the interface of that software.
You program your simulation or upload your data and then hundreds of processors spring into action for a short burst of time to crunch the numbers. Seconds or perhaps minutes later the results of your simulation are available.
Examples Of Cloud-Based Simulations
Cloud-based simulation services are taking off and it seems that every few months we hear of a new service. While which service is potentially right for you is outside the scope of this article, there are a few prime examples that do a good job of showcasing the variety and breadth of this new technology application.
SimScale, pictured above, is used by engineers, scientists and designers. From a virtual wind tunnel to fluid dynamics and mechanical stress tests. You can do it all from a tablet, with their server farms doing all the heavy lifting.
AutoDesk offers cloud-based versions of its industry standard CAD software. CAD workstations are a major category of professional computer, but with a cloud solution you don’t have to worry about, for example, owning an expensive workstation GPU.
Both Ansys and Rescale offer HPC (high-performance computing) services through the cloud. It’s like renting your own supercomputer, just as long as you need it to do a specific job.
This is just the tip of the iceberg of the incredibly computing power that could soon be available to almost anyone.
The Pros Of a Local Workstation
Using a cloud-based simulation solution instead of using your own local professional computer is not yet the best idea for everyone. Indeed, right now it’s actually best for only a few key types of professional user to make the leap to the cloud.
So why would you want to stick to a local workstation? The first important advantage that this approach has is data privacy. If you use any sort of third-party computing facility, your data has gone out of your hands. So if the work is being done on something that’s very sensitive, you’ll have to be 100% sure that the security measures and privacy policy of the service provider is up to snuff.
The other big advantage of the local workstation option is that you can run a vast selection of software on it. Workstations have been the dominant solution for decades. So most of the support is still there. The fact is that your niche professional software package might not yet be on the cloud.
The Pros Of Cloud Simulation
When it comes to going the cloud simulation route, there’s a lot to be said for this novel new method of solving hard problems with hot computer hardware.
For one thing, cloud-based simulation doesn’t have a big upfront cost. You only pay for what you use or a fixed-subscription that includes a certain amount of usage. Either way, costs can be minimized. Think of it like electricity. You flip a switch and are billed for the power that you consume.
When you use local computing power, you have to pay for all the hardware, upkeep and upgrades while not using the machine 100% most of the time and perhaps needing it to provide more than 100% at critical times.
Mobility is another large factor. Virtually any device with an internet connection and modern web browser can make use of a cloud-based service. This means you don’t have to lug a big workstation-class laptop or portable desktop solution around. It’s pretty amazing that an engineer can whip out an iPad on-site and run simulations at will.
The Cloud Is Not For Everyone
There are of course downsides as well. For example, the reliability of the service is completely in the hands of the service provider which makes good service level agreements essential. For individual users, it means that you have to plan for the inevitable outages. Even if a service has 99% uptime, it will go down at some point.
You also need some pretty beefy internet bandwidth to move that data around and trust the company will protect it once in their hands. If these don’t seem like issues to you, then instead of upgrading those workstations, it might be time to sign on the dotted line.