Can someone please explain how cloud hosting works?
I've seen this diagram that explains it well.
++++++++++++++
list of top cheapest host http://Listfreetop.pw
Top 200 best traffic exchange sites http://Listfreetop.pw/surf
free link exchange sites list http://Listfreetop.pw/links
list of top ptc sites
list of top ptp sites
Listfreetop.pw
Listfreetop.pw
+++++++++++++++
https://i.imgur.com/JV2vmt3.jpg
But my question is, is my website stored on one physical server, and primarily uses the resources from that original server, but when needed, it will use CPU, ram and other resourses from other machines in it's cloud network? (I'm with HawkHost)
Or, is my website stored on one server, but uses the CPU, ram etc from other machines within its network?
And cloud storage, is that pretty much the same but uses the same type of technology but for storage only?
My understanding is that it's a cluster of hardware that behaves as a single server, so in that case no your data is still in one physical location. To have data in multiple locations is called "anycast" and involves routing different input connections to physically different servers located around the world.
It is more like a marketing trick. Cloud hosting is almost the same as regular hosting, some providers call themselves cloud, but they are not even like a cloud. As I understand, regular hosting accounts are built from separate servers, while cloud hosting accounts from clusters. Example:
Webhosting - provider takes one physical server and creates 100 webhosting accounts on it. These 100 accounts share same server's resources. When provider sells all the spots, it takes 2nd physical machine and creates another 100 accounts. First 100 and second 100 accounts are not related together, as they are on different physical machines.
Cloud hosting - provider takes 10 physical machines as a cluster, installs IaaS and creates 1000. These 1000 accounts share all the resources these 10 machines have.
The problem with the term cloud is it's interpreted very differently.
There are providers who consider a single server with CloudLinux as "Cloud Hosting".
There are providers who offer "Grid hosting" where the storage, web server, database is separated onto their own servers as "Cloud".
True cloud hosting is about high availability and extreme resilience to the level which nearly nobody can actually achieve, for example I would take the view that it isn't full cloud unless you can wipe out an entire continent (which has servers serving your service/website) and not see an interruption in service because it is automatically routed to another region.
The only providers who can really achieve this to some level is Amazon, Google, Microsoft - but even so, 99.9999999999999% of people who base their service on these platforms are still not achieving that level of resilience as most only use one region, or don't have suitable failover configurations.
Basically, if you're talking about a standard website - don't worry about the cloud label. Focus on speed, reliability & customer service
Dan Rodgers - Managing Director - allthe.domains
800+ Domain Extensions Supported | Nominet & CentralNic Accredited Registrar | SuperFast SSD Cloud Hosting | 24x7x365 UK-based Support
The problem with the term cloud is it's interpreted very differently.
...
There are providers who offer "Grid hosting" where the storage, web server, database is separated onto their own servers as "Cloud".
True cloud hosting is about high availability and extreme resilience to the level which nearly nobody can actually achieve, for example I would take the view that it isn't full cloud unless you can wipe out an entire continent (which has servers serving your service/website) and not see an interruption in service because it is automatically routed to another region.
The only providers who can really achieve this to some level is Amazon, Google, Microsoft - but even so, 99.9999999999999% of people who base their service on these platforms are still not achieving that level of resilience as most only use one region, or don't have suitable failover configurations.
Basically, if you're talking about a standard website - don't worry about the cloud label. Focus on speed, reliability & customer service
+1
This has everything you need to get an idea of cloud. Thank you for describing the concept so smoothly.
Cloud can mean different things. It can be any of the following and does not have to mean all of them: HA (high availability), better scalability, billing per minute/hour, a better guarantee of not losing data due to failure on one single server (data is duplicated to 3 different storage devices). The list goes on and on.
People use Cloud hosting or Cloud VPS, Cloud Servers for different things. I use Cloud Servers because it allows me to upgrade the resources on my own within control panel unlike you can with a dedicated server where someone needs to take out the server and swap/update things physically. That said, dedicated servers have their pros too, and Cloud has their cons. Also, with Cloud Servers I like how I can add/remove attached drives within seconds as opposed to waiting for the datacentre to make such changes. I also like how the data on attached drives are duplicated to 3 different storage units.
Unfortunately, you will have some people who are biased and do not offer Cloud and only offer Dedicated servers and so will put Cloud down just because they do not or cannot provide it for various reasons. On the other side, there are some who provide Cloud and swear by it and will put dedicated servers down. The truth is, both have their uses and pros/cons.
Ps running CloudLinux on a single server is not Cloud hosting in a sense it is supposed to mean. CloudLinux with use of CageFS/LVE manager isolates accounts from one another for security and performance reasons only. You could say it is Cloud because it can improve uptime. But yeah, don't say you offer "Cloud hosting" just because you are running CloudLinux. Those who run a proper Cloud setup will run CloudLinux too.
As mentioned by others, there is no one definition of "Cloud". In the beginning "The Cloud" just meant a server on the internet. Now it can mean 1000 different things based on who you are talking to, etc.
The diagram link is a "Shared Cluster", but it still has its issues. In this setup you are assigned a set number of cores and ram that your website / application can use. There are many servers in the cluster that will serve the website, many more that might serve the database, etc. To keep it super simple, lets use a 1:1 ratio. Lets say each visitor to your site uses 1 CPU Core and 1GB of Ram. I know this is highly unreasonable, but like I said lets keep it simple. So if you purchase 2 CPU Cores and 2GB of Ram, then you can have two visitors on web site at the exact same time. Visitor A, might be using 1 Core and 1GB of ram from Server #1, where visitor 2 might be using 1 Core and 1GB of ram from Server #6. When visitor A leaves, and Visitor C visits the website, they might be using 1 Core and 1GB from Server #3.
All your data is stored on a central server(s) and each one of the web servers call back to that central location. Normally with companies like Amazon, Google, Azure have an outage, its the server cluster has lost connection to the data server, so it can not pull up any of your files. This is not 100% the case, but its the most common.
Again this is keeping it super simple, but that's the basis of how a Server Cluster. Technically you can create the same thing with a single server and setting up multiple VPSes. But doing so has no real benefit. Instead what is common is to create a private cloud with a couple of servers and do basically what the big boys are doing, but you have complete control over everything and often way more flexibility.
Tony Holloway
Delivering Custom Hosting Solutions Since January 2004
madRooster.com
Fun Name ... Serious Solutions
Technically you can create the same thing with a single server and setting up multiple VPSes. But doing so has no real benefit. Instead what is common is to create a private cloud with a couple of servers and do basically what the big boys are doing, but you have complete control over everything and often way more flexibility.
True. The big boys can offer "public" from their "private" setup". Some offer "public" only. So OP can learn the difference between "private" and "public" too.
Fast Host - Cloud Hosting w/ LiteSpeed & Imunify360 Security
Backup Storage - Backup Storage w/ SFTP, SSH & cPanel Access
NIST provides a fairly reasonable definition: https://www.technologyreview.com/s/4...uting-defined/
Dan Rodgers - Managing Director - allthe.domains
800+ Domain Extensions Supported | Nominet & CentralNic Accredited Registrar | SuperFast SSD Cloud Hosting | 24x7x365 UK-based Support
Two terms: redundancy and scaleability.
Redundancy: (at least) one server failure should not cause problem since everything is copied to at least one more place.
Scaleability: if you need more, or fewer resources, it can be achieved without any service interruptions - adding more, or fewer resources ("servers") to your "allotted cluster".
Not providing any of the above noted criteria won't stop every marketing department from advertising "cloud hosting", I'm sure.
Mostly harmless?
I/O Gremlin
j hostel kandy
cannesinfo.unblog.fr
clubedeautores.com.br
hostdime.com
imgdino.com
domain 0.5$
6 hostages in iran
domain buy service
hosting prices
url2it.com
hosting newark de
I've seen this diagram that explains it well.
https://i.imgur.com/JV2vmt3.jpg
But my question is, is my website stored on one physical server, and primarily uses the resources from that original server, but when needed, it will use CPU, ram and other resourses from other machines in it's cloud network? (I'm with HawkHost)
Or, is my website stored on one server, but uses the CPU, ram etc from other machines within its network?
And cloud storage, is that pretty much the same but uses the same type of technology but for storage only?
This is a fun topic that should be sticky'd I think!
The term "cloud" is ubiquitous and different providers have different definitions. In our case (Hawk Host) when we say "cloud" it means this:
All of our data is on a redundant SAN in the facility. We utilize Storpool for our SAN. [1]
All of our hypervisors (read: servers) exist on machines that have hefty CPU and memory.
If a hypervisor goes down (IE: a hardware failure entirely and unexpectedly) it'd be effectively the same as a physical machine crashing. The difference is our platform will auto migrate the customers to a brand new hypervisor and since the data is on a SAN the customer doesn't need to worry about the offline server. [2]
If a hypervisor is overloaded (IE: the usage is too high) we can hot migrate them to a brand new server alleviating any issues without any downtime.
[1] If something hard crashes there may be some nuance here.
[2] Our cloud web hosting backups are taken to a backup server outside of the SAN and to a different location.
One more thing to mention - pay as you go for system resources. For example: you need more CPU's for 12 hours, you pay only for 12 hours, you need more storage for a week - you pay for a week only. Of course it depends on billing system and does not require cloud to implement that, but the cloud is all about simplicity of scalability.
https://s3.amazonaws.com/awesomescre...i7%2BmrY7ps%3D
It seems to me that true cloud hosting would be similar to the term its referencing, 'cloud,' as in 'up in the sky,' or 'not connected or linked to the ground.' The 'connected/linked to the ground' part is where the Single Point of Failure comes into play, at least metaphorically, as that connection is required for functionality. It's a link that locks you into a a single connection or a single service and that is a vulnerability.
Quote Originally Posted by Aractus View Post
My understanding is that it's a cluster of hardware that behaves as a single server, so in that case no your data is still in one physical location. To have data in multiple locations is called "anycast" and involves routing different input connections to physically different servers located around the world.
Yes, without this your cluster is confined to one location and you have to rely on multiple network uplinks and distributed SANs for redundancy, but if this area is effected by a catastrophe, and the SANs and N+x architecture is destroyed, the data is no longer accessible. At this point you went from Cloud to Puddle.
Quote Originally Posted by DanDomains View Post
There are providers who offer "Grid hosting" where the storage, web server, database is separated onto their own servers as "Cloud".
If there is more than one web server, DB, and storage server and they don't rely on a single connection, this might actually work. With at least N+1 architecture for each service as a separate entity would make the cut. The issue at hand though with this, is that this is similar to what Aractus was saying. These would all have to be in separate locations so a catastrophe wouldn't effect the availability. Plus, there's failover time, which effects your nines (.9999)
Quote Originally Posted by DanDomains View Post
True cloud hosting is about high availability and extreme resilience to the level which nearly nobody can actually achieve, for example I would take the view that it isn't full cloud unless you can wipe out an entire continent (which has servers serving your service/website) and not see an interruption in service because it is automatically routed to another region.
This is also exactly right, at least if designed that way. The space between these "locations" might not have to be spread out that far. But you should be able at least to take a couple meteor hits without losing availability of your data.
Quote Originally Posted by DanDomains View Post
The only providers who can really achieve this to some level is Amazon, Google, Microsoft - but even so, 99.9999999999999% of people who base their service on these platforms are still not achieving that level of resilience as most only use one region, or don't have suitable failover configurations.
Yes solving this issue is where the money comes in, at least if people need a higher level of hosting.
Quote Originally Posted by DanDomains View Post
Basically, if you're talking about a standard website - don't worry about the cloud label. Focus on speed, reliability & customer service
The importance of true cloud hosting will always only be relevant to the owner of the data and not everyone needs high uptime, even though they should aim for it.
Quote Originally Posted by HostXNow_Chris View Post
Cloud can mean different things. It can be any of the following and does not have to mean all of them: HA (high availability), better scalability, billing per minute/hour, a better guarantee of not losing data due to failure on one single server (data is duplicated to 3 different storage devices). The list goes on and on.
HA and Scalability are definite musts.
Quote Originally Posted by HostXNow_Chris View Post
People use Cloud hosting or Cloud VPS, Cloud Servers for different things. I use Cloud Servers because it allows me to upgrade the resources on my own within control panel unlike you can with a dedicated server where someone needs to take out the server and swap/update things physically. That said, dedicated servers have their pros too, and Cloud has their cons. Also, with Cloud Servers I like how I can add/remove attached drives within seconds as opposed to waiting for the datacentre to make such changes. I also like how the data on attached drives are duplicated to 3 different storage units.
Yes I imagine Cloud servers are preferred as the client has no dealings with hardware issues, or they can at least leave them to the Cloud Server provider. Then you have to choose to suffer the limits of virtualization, or to suffer the limits of component failure, or both depending on VM Host node circumstances. At the end of the day, it's still a VM, sitting on baremetal box, possibly with N+x architecture, yet still vulnerable to a disaster knocking it off the map. It sucks having to balance these but luckily with HA setups the limitations of both types of environments are lessened.
Quote Originally Posted by HostXNow_Chris View Post
Unfortunately, you will have some people who are biased and do not offer Cloud and only offer Dedicated servers and so will put Cloud down just because they do not or cannot provide it for various reasons. On the other side, there are some who provide Cloud and swear by it and will put dedicated servers down. The truth is, both have their uses and pros/cons.
No matter what they choose, single points of failure will eventually catch up to them at some point. It's kinda like Murphy's law I'd imagine.
Quote Originally Posted by HostXNow_Chris View Post
Ps running CloudLinux on a single server is not Cloud hosting in a sense it is supposed to mean. CloudLinux with use of CageFS/LVE manager isolates accounts from one another for security and performance reasons only. You could say it is Cloud because it can improve uptime. But yeah, don't say you offer "Cloud hosting" just because you are running CloudLinux. Those who run a proper Cloud setup will run CloudLinux too.
Igor must be rich by now
Quote Originally Posted by madRoosterTony View Post
All your data is stored on a central server(s) and each one of the web servers call back to that central location. Normally with companies like Amazon, Google, Azure have an outage, its the server cluster has lost connection to the data server, so it can not pull up any of your files. This is not 100% the case, but its the most common.
You'd think these companies would would have multiple data servers so this wouldn't happen.
Quote Originally Posted by madRoosterTony View Post
Again this is keeping it super simple, but that's the basis of how a Server Cluster. Technically you can create the same thing with a single server and setting up multiple VPSes. But doing so has no real benefit.
Yeah same location would be pointless and have an equivalent to a SAN setup.
Quote Originally Posted by DanDomains View Post
NIST provides a fairly reasonable definition: https://www.technologyreview.com/s/4...uting-defined/
This is old, but I agree true cloud providers should be at least performing at this level --> "Software as a Service (SaaS) is at the top of the cloud computing stack. Here the cloud providers have created full applications running on server farms that may themselves be geographically distributed. " If it can fail because it's somehow limited to a single point of failure on Earth, it really doesn't represent a cloud.
Quote Originally Posted by CodyRo View Post
This is a fun topic that should be sticky'd I think!
The term "cloud" is ubiquitous and different providers have different definitions. In our case (Hawk Host) when we say "cloud" it means this:
All of our data is on a redundant SAN in the facility. We utilize Storpool for our SAN. [1]
All of our hypervisors (read: servers) exist on machines that have hefty CPU and memory.
If a hypervisor goes down (IE: a hardware failure entirely and unexpectedly) it'd be effectively the same as a physical machine crashing. The difference is our platform will auto migrate the customers to a brand new hypervisor and since the data is on a SAN the customer doesn't need to worry about the offline server. [2]
If a hypervisor is overloaded (IE: the usage is too high) we can hot migrate them to a brand new server alleviating any issues without any downtime.
Storpool is a great solution. But then again it's limited to the "facility."
Quote Originally Posted by bikegremlin View Post
Two terms: redundancy and scaleability.
Redundancy: (at least) one server failure should not cause problem since everything is copied to at least one more place.
Scaleability: if you need more, or fewer resources, it can be achieved without any service interruptions - adding more, or fewer resources ("servers") to your "allotted cluster".
This is exactly correct and how it should be.
Quote Originally Posted by bikegremlin View Post
Not providing any of the above noted criteria won't stop every marketing department from advertising "cloud hosting", I'm sure.
This is for sure. This is how we ended up where we are now with the 300 definitions of what "cloud hosting" is.
Google, the king of search engines is itself a good example of real-time cloud hosting. It has got its resources spread over hundreds of servers on the cloud, no wonder you've never seen Google.com facing any downtimes over past decade or so.
Each server in the cloud helps in carrying out a particular set of tasks, and in the case of failure of any of the servers in the cloud, another server temporarily kick-in as a back-up.
I've seen this diagram that explains it well.
https://i.imgur.com/JV2vmt3.jpg
But my question is, is my website stored on one physical server, and primarily uses the resources from that original server, but when needed, it will use CPU, ram and other resourses from other machines in it's cloud network? (I'm with HawkHost)
Or, is my website stored on one server, but uses the CPU, ram etc from other machines within its network?
And cloud storage, is that pretty much the same but uses the same type of technology but for storage only?
The main principle of cloud hosting is geo-distribution and at least double redundancy.
Data is written to several servers at once that overlap, often in different data centers.
Some very informative answers here - thanks for taking the time to write these!
If you need to host an application or website that may need to scale cloud is a good solution. The storage is going to be on a SAN not hard drives on a dedicated server. This gives you the ability to increase storage in a production environment. It is also the reason storage on cloud cost so much. The processor and RAM live on a host. If that host fails your site will migrate to a different host. An example of this would be VMware's vMotion or Microsoft's Cluster Server (MSCS). Proxmox automatic failover is another example among many examples. Hope this helps.
https://cari.net/
Building custom hosting solutions for over 15 years!
Email Us: salesgroup [at] CARI.net| Facebook: https://www.facebook.com/CARInetInc | Twitter: https://twitter.com/carinetinc
Cloud Hosting does not mean your Operating system is running on Multiple Systems at the same time ( read processor+Ram) .It simply means your server/system runs on a group of systems wherein your infrastructure can be expanded (add more ram/cpu) without significant downtime and the Physical issue affecting a Physical server may not impact you because you can live migrate your instance to another physical host . Most people in this forum will relate a cloud as IaaS because they are all Infrastructure providers , while others can also mean something like SaaS ( Google Docs etc) or a PaaS ( RedHat OpenShift platform)
A U T O M 8 N . C O M
cPanel plugin for Active-Active redundancy,High Availability and Native Nginx
DDOS-BruteForce-BadBot mitigation, System Monitoring using Netdata, cPanel Backup using Borg/Borgmatic
User Friendly,Automated! ,Read the Docs -- https://autom8n.com/docs/
