Web Hosting Talk : Web Hosting Main Forums : Cloud Hosting : How are resources managed in cloud hosting?

[torwill]

I kind of understand what cloud is after reading some articles. But I can not quite figure out how resources are managed in cloud hosting?

Lets say I have a cloud of 2 servers, each has one quad core cpu, 8gb ram and 200gb local storage devoted for vm. please also presume there is no overhead on hypersor servers.

Now, I have a VM that requires 6 core of processing power, 12gb of ram, and 300gb of storage.

Would the clound controller automatically spread tasks on this VM to two servers? (4 core/8gb ram/200gb storage on server1, 2core/4gb/100gb storage on server2)

Is this how cloud works?


Thank you.

[anthonysomerset]

no because what that would be is distributed processing

if you wanted a 6 core, 12gb ram, 300gb hd vps/vm you would need a hypervisor that was at list that big to be able to have it

the cloud controller depending on what you use only assigns vm's to servers and manages them from the outside (power network etc)

[Winky]

Originally Posted by torwill

I kind of understand what cloud is after reading some articles. But I can not quite figure out how resources are managed in cloud hosting?

Lets say I have a cloud of 2 servers, each has one quad core cpu, 8gb ram and 200gb local storage devoted for vm. please also presume there is no overhead on hypersor servers.

Now, I have a VM that requires 6 core of processing power, 12gb of ram, and 300gb of storage.

Would the clound controller automatically spread tasks on this VM to two servers? (4 core/8gb ram/200gb storage on server1, 2core/4gb/100gb storage on server2)

Is this how cloud works?


Thank you.

Cloud doesn't do that by default. Think of cloud instances as dedicated servers in the way they function. Your two servers are two separate systems, and they won't just merge together at will. If you need a server with 6 cores and 12GM RAM, you'll need to provision one of the servers to have that much.

If you're using VMware, there is something called DRS (distributed resource scheduling). What that does is, if a VM is on a host that has low resources, the software will migrate the VM to another host that has the resources available using vMotion technology. So, DRS ensures that a VM has the resources it needs if the current host it's on can't provide it. But this is different than a VM gaining additional RAM or CPU in real time to handle a process.

[CloudWeb]

Also, if you only have two servers in a Cloud then you should only ever use the resources of one. Why? What happens when one fails? you will have no redundancy or high-availability as you need enough free resources to run all of the applications on your cloud.

While a Cloud can technically be built with just two servers, it's obviously inefficient at this level.. with 2 servers you lose 1 that's a 50% reduction, with 10 servers you lose 1 it's only a 10% reduction and so on...