Howdy, Stranger!

It looks like you're new here. If you want to get involved, click one of these buttons!


Real Defination of Cloud Hosting - Page 2
New on LowEndTalk? Please Register and read our Community Rules.

All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.

Real Defination of Cloud Hosting

2»

Comments

  • hostbethostbet Member, Host Rep

    The defination for cloud is different for everyone here.

    If you are talking about cloud computing then it's basically a concept in which we connect multiple servers to prevent our application going down but most of provider doesn't follow this concept.

    I believe it is more likely a marketing word, i have seen many companies saying Cloud Hosting Cloud VPS but they really doesn't mean that they have multiple servers connected together and your application will not go down. a minority of providers have implemented the real (actual) cloud computing, rest are just marketing.

  • jarjar Patron Provider, Top Host, Veteran
    edited November 2021

    @hostbet said:
    The defination for cloud is different for everyone here.

    Right, because some people make up another meaning out of thin air. Like I said, hosts that feel like they can't offer cloud hosting are making it up to downplay the value of competition. How fitting that they choose a definition that allows them to declare all of the big clouds "not real cloud" so as to say "they don't offer anything the rest of us don't, just marketing words." Sure enough, guess who pops up to argue that made up definition against facts...another web host that doesn't have a cloud offering. You know the big clouds do a lot of things you don't, but you don't want people to see that so you lie to them.

    No reputable and well known company is selling hosting they call "cloud" that doesn't see their product as fitting some part of the NIST definition of cloud computing. Show me one.

    I believe it is more likely a marketing word, i have seen many companies saying Cloud Hosting Cloud VPS but they really doesn't mean that they have multiple servers connected together and your application will not go down. a minority of providers have implemented the real (actual) cloud computing, rest are just marketing.

    This isn't religion, what you believe doesn't matter. You just took the wrong definition and ran with it like it was fact. You can't just make up definitions for words and then declare everyone else to not meet the definition of the word. I smell exactly what I previously said, a web host that wants to downplay the value of their perceived competition. On demand resources, hourly billing, you feel like you can't compete so you try to twist reality.

    Notice the key:

    i have seen many companies saying Cloud Hosting Cloud VPS but they really doesn't mean that they have multiple servers connected together

    The acrobatics here are nice. Why would it matter that they say they have cloud VPS and then not provide something that doesn't meet the real definition of cloud computing? Because you just wormed your way into declaring the false meaning to be the real one. Nice play, HostBet. Dishonesty won't get you sales here.

    If you don't think you stand to gain from misinforming people, I can't see another reason to be married to the idea of doing it. To recognize the truth, acknowledge it, and then continue to misinform. That's downright malicious.

    And now that I don't represent a cloud provider, I get to call people like you out for your shady marketing tricks. Feels good man. Legacy web hosts are such slimy people. That's why I had to ban nearly everyone who signed up during HostingCon one year, after a presentation that mentioned LET as a place to market hosting. Those types are insufferable, incompetent, and belong on WHT. Congrats on your 5th post though. I liked how you made two comments just to get there. Did you get your ticket in to request the provider tag?

  • bikegremlinbikegremlin Member
    edited November 2021

    @jar great input - much appreciated.

    I was taught that cloud requires both scalability and redundancy in order to be called that.
    Was that wrong, and is the Wiki article more correct? Possibly.
    Though I trust your knowledge and experience more than a Wiki article. :)

    Is it OK to quote you in the article (when I edit the info)?

    Thanked by 1jar
  • bulbasaurbulbasaur Member
    edited November 2021

    @bikegremlin said: I was taught that cloud requires both scalability and redundancy in order to be called that.

    Scalability and redundancy come from the fact that you have the ability to request resources immediately upon request. As an example, if a server fails, you can destroy it and spin up a new one with the exact same configuration as the previous one.

    Some of the bigger clouds automatically do this for you (AWS Autoscaling groups comes to mind), but I wouldn't consider it a necessary condition as most people would consider Vultr or DigitalOcean as "clouds" even though only the primitives to automatically provision resources exist in the form of API calls, without these being managed for you like AWS does.

    @bikegremlin said: Though I trust your knowledge and experience more than a Wiki article. :)

    The definition on Wikipedia comes from NIST though.

    Thanked by 2bikegremlin jar
  • @stevewatson301 said:

    @bikegremlin said: I was taught that cloud requires both scalability and redundancy in order to be called that.

    Scalability and redundancy come from the fact that you have the ability to request resources immediately upon request. As an example, if a server fails, you can destroy it and spin up a new one with the exact same configuration as the previous one.

    Some of the bigger clouds automatically do this for you (AWS Autoscaling groups comes to mind), but I wouldn't consider it a necessary condition as most people would consider Vultr or DigitalOcean as "clouds" even though only the primitives to automatically provision resources exist in the form of API calls, without these being managed for you like AWS does.

    @bikegremlin said: Though I trust your knowledge and experience more than a Wiki article. :)

    The definition on Wikipedia comes from NIST though.

    It's been my understanding that the ability to "survive" at least one physical machine failure was one of the pre-requisites for being called "cloud."

  • bulbasaurbulbasaur Member
    edited November 2021

    @bikegremlin said: It's been my understanding that the ability to "survive" at least one physical machine failure was one of the pre-requisites for being called "cloud."

    This can only happen under limited circumstances where the provided service is stateless.

    1. A storage service like AWS S3 or Google Cloud Storage can work around node failures in a similar way RAID does; data can be recovered from different node in case of failure.

    2. A serverless function service such as AWS Lambda, Google Cloud Functions or Cloudflare functions (basically, you upload a script and it is triggered by HTTP requests) is inherently stateless, since the code is packaged as a single unit which can be run in a standard environment.

    With a server or database, its state could potentially be always changing and hence stateful; thus synchronization always needs to be handled at the application level. Even the big clouds can't handle that; if an instance has a hardware failure AWS asks you to stop and start the instance or in the case of their database service, recover from a backup.

  • @stevewatson301 said:

    @bikegremlin said: It's been my understanding that the ability to "survive" at least one physical machine failure was one of the pre-requisites for being called "cloud."

    This can only happen under limited circumstances where the provided service is stateless.

    1. A storage service like AWS S3 or Google Cloud Storage can work around node failures in a similar way RAID does; data can be recovered from different node in case of failure.

    2. A serverless function service such as AWS Lambda, Google Cloud Functions or Cloudflare functions (basically, you upload a script and it is triggered by HTTP requests) is inherently stateless, since the code is packaged as a single unit which can be run in a standard environment.

    With a server or database, its state could potentially be always changing and hence stateful; thus synchronization always needs to be handled at the application level. Even the big clouds can't handle that; if an instance has a hardware failure AWS asks you to stop and start the instance or recover from a backup (because they inherently can't because of the stateful nature of these services):

    https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/monitoring-system-instance-status-check.html
    https://aws.amazon.com/premiumsupport/knowledge-center/rds-instance-failed-state/

    Just to confirm I understood correctly - apologies in advance if it's a stupid question:

    With "non-cloud" servers, one HDD/SSD failing can still let the machine to be up and running if a proper RAID is configured. But motherboard or a CPU breaking down will cause the server to go down.

    "Cloud servers" use virtualization. One virtual machine ("cloud server") can use more than one physical server to run, just as one physical server can run several virtual "cloud servers." Right?

    Provided I got that right - the machine can be called "cloud" regardless of whether it can stay up and running in case either a storage unit (HDD/SSD) fails, or a motherboard/CPU fail?

  • @bikegremlin You'd want to look at the NIST definition of cloud, it's all about being able to rapidly request for resources (compute, storage or otherwise) with minimal effort. Hourly billing, provisioning through APIs and minimal intervention by the requester or provider fall under that; which is why "cloud" services are called such.

    All providers, even non-cloud ones, use virtualization, except for dedis, where by definition you can't have virtualization. Although, a single virtual machine typically does not run across multiple machines; you'll only find that stuff in supercomputers or HPC clusters that have a framework to carry compute bound tasks around to different nodes.

    Thanked by 2bikegremlin jar
  • FractionFrankFractionFrank Member, Host Rep

    I don't think there will ever be a single definition of cloud - unfortunately for the consumer! Some less technically-minded people often refer to the "cloud" as just being a server or end-point away from their own premises.

    I think a true "Cloud" solution is one that has redundancy in the form of multiple hypervisors and storage nodes at minimum.

  • jarjar Patron Provider, Top Host, Veteran
    edited November 2021

    @FractionFrank said:
    I don't think there will ever be a single definition of cloud - unfortunately for the consumer! Some less technically-minded people often refer to the "cloud" as just being a server or end-point away from their own premises.

    I think a true "Cloud" solution is one that has redundancy in the form of multiple hypervisors and storage nodes at minimum.

    It does have a single definition though. Thanks for your input, another legacy web host. We know, all of the big clouds are just marketing fluff and don't offer anything that FractionServers doesn't! 🙄

  • jarjar Patron Provider, Top Host, Veteran
    edited November 2021

    @bikegremlin said:
    @jar great input - much appreciated.

    I was taught that cloud requires both scalability and redundancy in order to be called that.
    Was that wrong, and is the Wiki article more correct? Possibly.
    Though I trust your knowledge and experience more than a Wiki article. :)

    Is it OK to quote you in the article (when I edit the info)?

    You can but it's better to quote the article. It's very well written and pulls from reputable sources.

    The bonus is that once you use that definition you find that AWS, Azure, DigitalOcean, Vultr, and all similar cloud providers are perfectly aligned with the definition. When the definition and reality match up, it's a nice fit ♥️

    And yeah, the ability to spin up resources on the fly without the provider's involvement is key to making fully self healing, redundant, and auto scaling applications. The problem is teaching users that this is already done for them when they just spin up a VM on a "true cloud provider" because that largely doesn't exist in such a simple fashion. You don't want to use a definition that consistently doesn't fit reality, it's not fair to end users who lose all of their data and then start crying "but you said it was cloud." Sadly, I've had to talk those people off of a ledge too many times.

    Thanked by 1bikegremlin
Sign In or Register to comment.