Talk:Serverless computing

About merging serverless computing with Function as a Service

(merging Serverless computing with Function as a Service)

I object. According to [1], serverless encompasses FaaS and mobile backend as a service. Admittedly I have seen other articles where serverless is a synonym for FaaS. JnRouvignac (talk) 08:51, 31 July 2016 (UTC)

Mildly object. I think that serverless computing is broader than just FaaS and should include PaaS, so these aren't the same thing. The FaaS article should be more widely developed since there is plenty more to say about it. StacksofHoy (talk) 12:49, 15 August 2016 (UTC)

It's just such a big topic, and so fast moving - I'd let them all have their own page. "traditional music" "folk music" - sure, lots of overlap but plenty to be said about each sense. I think the warning can be taken down and leave it as is for now - maybe in three years the terms will be more settled and one or two can be eliminated. 24.42.191.195 (talk) 20:16, 11 November 2016 (UTC)

I second the mild objection to FaaS and serverless being combined. Serverless is less about functions as payloads of executions, and more about not having to administer servers. --Huntc (talk) 05:40, 3 March 2017 (UTC)

Third the objection. I disagree that BaaS systems that require you to purchase fixed-price plans are serverless, but certain serverless does and will extend well beyond FaaS platforms. The hallmark of serverless to me is pay as you go/metered billing.

Note that "AWS Lambda, Google Cloud Functions and Microsoft Azure Functions" appear in both articles as the manifestation of the title. this says something. — Preceding unsigned comment added by 185.127.10.42 (talk) 13:48, 2 July 2017 (UTC)

We now see "serverless apps" cited as a top security issue [2], and so it might be helpful to have a direct Wiki reference to this, even if it redirects to severless computing.CeleryStickz (talk) 17:59, 7 December 2017 (UTC)

I may be beating a dead horse at this point, but I also object. With the evolution in the topic since, it has become clear that the two are distinct. Serverless computing centers around an architectural design pattern of utilizing managed services, where FaaS is potentially one of the managed service offerings. That's to say, while a serverless application may include managed functions (i.e. FaaS), it is not a strict requirement[3]. On a high level, managed services could be roughly broken into two categories: (1) managed compute resources and (2) managed backends. The former could then be further broken into managed functions (i.e. FaaS - e.g. AWS Lambda, GCP Functions, Azure Functions, et al) and managed containers (i.e. CaaS - e.g. AWS Fargate, Iron.io, perhaps others). The latter, managed backends, encompasses a wide variety of backend service offerings; things like storage, databases, queues, caches, authentication, etc. - HarryLarold (talk) 22:49, 21 September 2019 (UTC)

Article quality

The quality of this article is abysmal. It seems to be an assessment of current offerings and a limited comparison of their features while avoiding using specific names of platforms. I particularly object to the assessment that FaaS is inherently insecure because it is FaaS. I'm currently looking for a FaaS solution with structured role based access control on a per-function basic. I also object to the assessment of performance related to the time required to start engines. Performance on systems based on some technologies may suffer performance issues related to the reasoning mentioned, however other systems which employ AoT compilers would have problems related to cold-start, other platforms might have the engines for the platform running at all times and simply suffer delays while obtaining the function itself from a document store. Finally, most performance issues could be avoided by compiling and caching functions on commit. I would fix the article myself, but have absolutely no idea where to begin as it is so poorly written I would need to perform proper research and start from scratch. Is it possible to simply revoke an article or label it as horrible? User:submux 11:17, 12 December 2017 (UTC)

@Submux: Have a look at WP:TNT, balanced by WP:BATHWATER. If you thought that a better article could be written from scratch, then one way to do it is to draft an article, perhaps in your sandbox, then launch a request for comment. Klbrain (talk) 14:31, 8 April 2018 (UTC)

The paragraph beginning "Find the balance where developers don’t own security…" consists of four sentences, all imperative. This strikes me as violating the neutrality principle - I can't remember seeing another imperative sentence anywhere in the wikipedia. The paragraph cites but does not indicate that that page's author is the source of the recommendations. I would fix this myself, but I know next to nothing about the subject. Tameware 02:11, 5 April 2019 (UTC) — Preceding unsigned comment added by Tameware (talk • contribs)

References

No Servers? No Servers?

There isn't anything in the article to explain how there are no servers. Is it peer to peer? How does it work without any DNS Servers, Web Servers or even Internet etc... completely Serverless? ZhuLien (talk) 18:00, 10 January 2018 (UTC) — Preceding unsigned comment added by 115.70.207.235 (talk)

There is now discussion of this 'misnomer' in the lede.

Resolved

Klbrain (talk) 11:19, 26 March 2018 (UTC)

I have added a bit stating serverless computing's name is a misnomer, as it requires servers (according to this article). I have also cited this twice using references.

And yes, in it's current form IMO, the name strikes of weasel words that only a cloud salesman would be happy using. 2A00:23C5:F089:FB00:E2CA:94FF:FEBC:20AF (talk) 18:32, 4 September 2020 (UTC)

The section certainly could use some improvement, but you need to provide a reliable expert source (not blogs or PR articles) for such content, especially for any kind of extraordinary claim. GermanJoe (talk) 22:11, 4 September 2020 (UTC)

I think even with good references, the sad fact is, it'll be taken down by people who wish to call ...

automatic cars "gearless" because you don't need to manually change gears.

The people who use this kind of terminology endorse snake oil.

81.151.152.185 (talk) 22:21, 6 September 2020 (UTC)

References

Privacy concerns

The privacy concerns listed in the article seem amateurish and apply to any hosted service. It also carries a certain amount of naivety about how cloud operators run PaaS services. FaaS are just a type of PaaS and 99.99% of respectable cloud operators do not give their employees access to the customer's payload and have safeguards against employees looking at customer's payloads willy-nilly. But again, the main criticism applies to any hosted service (IaaS, even just renting racks at Rackspace). — Preceding unsigned comment added by 2601:600:8780:2AE1:D90B:1F78:C182:2E6 (talk) 08:02, 8 July 2018 (UTC)

Promotional name dropping and link spam removed

Noteworthy details should be based on independent reliable sources. Wikipedia is no link directory or advertising platform for self-published PR fluff. GermanJoe (talk) 11:28, 11 July 2018 (UTC)

Unsubstantiated assertion about volatile memory and storage

In the article is the following text: "Serverless computing does not hold resources in volatile memory; computing is rather done in short bursts with the results persisted to storage." I'm not going to delete this without discussion, but it seems misguided to me. Surely the most obvious and primary meaning of "Serverless" is that we don't need to know anything about the implementation of the infrastructure or platform. A client calls a service/function and gets a result back. If the vendor chooses to use volatile memory or do something in short bursts, that's up to them. If they change it tomorrow to something entirely different, it will still serverless computing in the way we understand it. It's an abstraction. The intention behind this text is not clear to me. Just because it's incorrect/unsubstantiated doesn't mean there's not a valuable thought here that needs to be captured. What did the author intend to say?

Dominic Cronin (talk) 09:26, 5 December 2022 (UTC)

Tone

The intro is written with some elements like an instructional guide, which Wikipedia is not. That includes the use of "note that", "you", and "should". I've only read the intro myself, but I imagine the rest of the article contains similar verbiage. — W.andrea (talk) 16:54, 7 June 2024 (UTC)

Exception: "should not be confused with" is fine AFAIK. — W.andrea (talk) 17:06, 7 June 2024 (UTC)

I just moved some of the content to a lower section. Azarboon (talk) 01:15, 8 June 2024 (UTC)

I can't see "you" or "note that" in the page. Also, what's the alternative for "should"? Those containing sentences look neutral. Azarboon (talk) 01:18, 8 June 2024 (UTC)

I just changed the "should" with "advise" Azarboon (talk) 01:29, 8 June 2024 (UTC)

Someone already got "you" and "note that". — W.andrea (talk) 18:52, 8 June 2024 (UTC)

The alternative to "should" is explained in the link I posted:

Wording can easily be modified to avoid advising the reader: Do not give aspirin ... ⇒ The WHO advises against the use of aspirin ...

By the same token, "it is advised" is still too instructional, but you can just add attribution to the advice so that it's not in Wikipedia's voice, instead we're just repeating what someone else said (like I wrote at Talk:Microservices § Tone). — W.andrea (talk) 18:59, 8 June 2024 (UTC)

Thanks for instructions. Changed it accordingly. Azarboon (talk) 00:53, 9 June 2024 (UTC)

Problematic "Serverless databases" section

While I agree that databases marketed towards usecases with ephemeral runtimes are often used in serverless computing usecases, I find this section weird.

Do we need to be listing commercial products offerings? Especially since these are not the most popular database providers as of 2024 (citation needed). This seems like an ad board, which does not belong on Wikipedia. Jonte from Jontes.Page (talk) 19:53, 9 October 2024 (UTC)