TE
科技回声
首页24小时热榜最新最佳问答展示工作
GitHubTwitter
首页

科技回声

基于 Next.js 构建的科技新闻平台,提供全球科技新闻和讨论内容。

GitHubTwitter

首页

首页最新最佳问答展示工作

资源链接

HackerNews API原版 HackerNewsNext.js

© 2025 科技回声. 版权所有。

Why The Future Of Software And Apps Is Serverless

36 点作者 hausburger超过 12 年前

11 条评论

dpark超过 12 年前
&#62; <i>The phrase “serverless” doesn’t mean servers are no longer involved.</i><p>So it's a meaningless buzzword. Got it.
评论 #4655979 未加载
评论 #4655929 未加载
orthecreedence超过 12 年前
Ok, so 20 years ago, we all had to rely on lowly, pathetic servers. But <i>now</i>, thanks to <i>THE CLOUUUUDD</i> (read: hourly pricing on servers) there is a nebulous sludge of buzzwords running your applications, instead.<p>Seriously, come on. People still have to think about their servers. AWS is nice, yes, <i>but it still uses servers</i>, and developers and ops people <i>still need to think about those servers</i>.<p>It not like I press the magic button and my distributed app suddenly becomes available for millions to use simultaneously. I have to think (really hard) about what services I need, how they interact, how they scale in relation to each other, how to deal with failure scenarios, etc etc. None of this has gone away with the advent of "cloud" computing, and there is no way to invisibly scale your app (save maybe Heroku-like services, but they break down at high load).<p>Is it becoming easier to deploy scalable applications? Yes. But let's not confuse the fact that hourly pricing on VPS instances is still just running your app on a VPS, no matter how many of them there are. You still need to know your fair share of unix commands and you still need to be smart about the different pieces and how they interact. <i>This is nothing new</i>.<p>Let's also not forget what the cloud really is: an annoyingly overused buzzword thrown around by people who have never actually logged into a server in their lives.
评论 #4657749 未加载
评论 #4656523 未加载
malandrew超过 12 年前
This article has the right title and discusses the idea of services, not servers, but misses an even bigger opportunity to discuss how you can legitimately ditch lots and lots of servers. WebRTC means that every application can have aspects of peering and reduce the need for centralization in servers.<p>Back in 2010, well before WebRTC, I posted the following use case on the WHAT-WG mailing list that explores using peering to offset CDN static asset serving costs in social games. It was a while ago when I was a product manager and a lot less technical, so excuse some of the naive misconceptions I may have made at the time. <a href="http://lists.w3.org/Archives/Public/public-whatwg-archive/2010Jan/0103.html" rel="nofollow">http://lists.w3.org/Archives/Public/public-whatwg-archive/20...</a><p>Anyone interested in exploring an idea like iron.io, but in their own application stack should check out the following projects from James Halliday (substack)<p><a href="https://github.com/substack/seaport" rel="nofollow">https://github.com/substack/seaport</a><p><a href="https://github.com/substack/airport" rel="nofollow">https://github.com/substack/airport</a><p><a href="https://github.com/substack/fleet" rel="nofollow">https://github.com/substack/fleet</a><p><a href="https://github.com/substack/airport-cluster-example" rel="nofollow">https://github.com/substack/airport-cluster-example</a><p>It'd be awesome to see someone combine WebRTC with the basic idea behind Seaports to allow semantic versioning of services provided by your application's users.
peterwwillis超过 12 年前
<i>The phrase “serverless” doesn’t mean servers are no longer involved. It simply means that developers no longer have to think that much about them. Computing resources get used as services without having to manage around physical capacities or limits.</i><p>Every application has to deal with capacities and limits. If they didn't, they'd break.<p>Say your app just stores a couple hundred megabytes every second, because you imagine there's no such thing as a storage limit. Let's imagine disk storage could be expanded fast enough to support this and costs zero money. We still live in a universe bound by physics, and storage only goes so fast: you will eventually have too much data to process.<p>We can't assume CPU or RAM capacity is limitless. If your app is cracking crypto passwords, <i>depending on the password and method</i>, all of AWS's collective compute cycles still might not be enough to crack one password in a reasonable amount of time. Ask the people trying to make flight price comparison engines about resource starvation.<p>It's not even difficult to learn how servers work and affect your application. There's really no point to this crap.
0wza超过 12 年前
Well, what if I don't need the "cloud" to run virtual machines? What if I have many "machines" running on my computer, isolated kernels in userspsace? What if I have all the computing power I need at my fingertips, literally? Is that sort of control and transparency not useful? How many more security breaches do we need to see before we acknowledge that delegating our storage and computing needs may not be the wisest option in all cases?<p>The only thing I expect from the "cloud" is proper routing.<p>All the rest of the storage and computing can be done at the endpoints. This isn't some revolutionary idea. It's how the internet was originally imagined by Paul Baran.<p>Endpoints will vary. They will have different needs and capacities to meet them. Some may be even be "services" to which other subscribe. But there's no requirement for ubiquitous middlemen. Everything is not a service that requires a third party.<p>What's next, TaaS? Thinking as a Service?<p>The future of software is one of empowerment with less third party involvement.
ilaksh超过 12 年前
The main point I think is that services with web APIs are taking the place of individually managed servers. Which is a good point and a trend because it does make sense a lot of the time.<p>From the title I thought he was going to go farther towards the peer-based content-centric ideas, where the network nodes being addressed are actually data nodes rather than server nodes, and the data is distributed by peers.<p><a href="http://en.wikipedia.org/wiki/Content-centric_networking" rel="nofollow">http://en.wikipedia.org/wiki/Content-centric_networking</a><p>Many dismiss that type of peer-based networking as something that will 'never' work for the mainstream internet, but I think you just need sophisticated encryption (bitcoin level) and good upstream bandwidth.
ajdecon超过 12 年前
In other words: PaaS or SaaS will become much more popular to build on, and developers will spend less time working with servers or IaaS. Industry trends continue. We need an article for this?<p>From a practical perspective, part of what this means is that even more seemingly-distinct services will depend on the same infrastructure. We're seeing this already, as every time AWS experiences downtime, a bunch of web services go down. It will be interesting to see what effects this has on everyday business, as outsourcing internal infrastructure (email, calendaring, file storage, etc) becomes even more popular.
mark_l_watson超过 12 年前
I have a problem with the terminology in the article. For me, a server and the 'cloud' both refer to remote computing resources, separate from clients. The Venn diagram overlaps a bit :-)
pedalpete超过 12 年前
This is just a pendulum that swings back and forth, with 'serverless' as a new buzzword.<p>Everything was on the server (mainframe) with dumb terminals, where the mainframe was essentially the cloud, then we went to fat-terminals with the pc-era. The web came along and everything went back to the server, then smartphones and tablets, and we went back to apps with data stored on the server. Now we're going back again.<p>Where is this all going?
pjmlp超过 12 年前
I didn't go it. After all I've been coding distributed applications since early 90's.<p>So what changed in network topologies?
评论 #4656169 未加载
评论 #4656338 未加载
mephi5t0超过 12 年前
The first thing popped in my head after reading the title: "Your mom is serverless"