LinuxQuestions.org

LinuxQuestions.org (/questions/)
-   Linux - General (https://www.linuxquestions.org/questions/linux-general-1/)
-   -   Cloud vs Virtualization (https://www.linuxquestions.org/questions/linux-general-1/cloud-vs-virtualization-775931/)

jt1 12-15-2009 08:16 PM

Cloud vs Virtualization
 
Hello,

I have done HOURS of research on this topic but I am still unable to understand the difference. The articles that I have been reading make these two technologies seem similar except that cloud is less secure. So what is the big deal about cloud?

Also, how would one set it up or does anyone know of a control panel that is not too expensive.

Thanks.

sundialsvcs 12-15-2009 08:20 PM

"Cloud" is a tom-swifty marketing term for distributed computing. Servers are located around the network, looking for work to do.

"Virtualization" is a technique for taking a single computer and making it appear to be running multiple "virtual machines," each one of which is running a (perhaps different) "guest operating system."

The two techniques are, in fact, unrelated.

jt1 12-15-2009 09:36 PM

Hello All,

So how could someone setup cloud computing or is there some control panel software that is not too expensive so I can use it in a test environment?

jt1 12-29-2009 10:30 PM

Is cloud computing similiar clustering?

Quakeboy02 12-30-2009 12:01 AM

What are you trying to accomplish? Are you doing research for homework or for some business need at work? Are you trying to run multiple operating systems on one computer? Do you have several computers you want to somehow hook together? Give us some idea of what you are looking for and maybe we can offer some real help.

jt1 12-30-2009 12:47 PM

Thanks for the reply.

I have an environment in which I have 5 servers. The servers have roughly 7 or 8 VMs. The os on all the servers are RHEL and the VMs comprise of various linux distros. I am using open source xen with each server. Three of the server are always full to capacity in terms of the processor and the memory. The other two are not so much. What I want to do is either build a cluster in which all of the servers act as one allowing me to be able provide "unlimited" resources to the VMs or have computers that are on standby and when the time is needed the servers in production could use the resources on the computers that are on stand by. I thought the latter part was cloud computing so I wanted to build myself an infrastructure similar to that.

I wanted to know if anyone could point me into the right direction on which technologies I can utilize. I researched Hadoop and it seems that it is mainly for storage. I want to build an environment that is completely scalable in terms of storage, memory, and processing power and still have the ability to add additional VMs and servers. If what I want to accomplish is not practical, please let me know.

Thanks again for any help.

Quakeboy02 12-30-2009 01:20 PM

I don't think there is a simple point and click answer for you. You have 3 machines out of 5 that are at capacity. Your choices seem limited to either getting larger machines, or developing a capacity management system where users are shunted off to other machines; either due to server usage level or a random(ish) assignment when logging in or tasks are being performed.

You also need to be sure that you don't simply have too much load on your DBMS (Oracle, whatever).

If I were faced with this problem, I'd sit down with some of my peers, and discuss the work that is actually being done, the hardware capacity, and the network topology. Look for ways to rearrange things to lighten the load. If you can't rearrange software, hardware, network, or database, then larger servers are in your future.

Added:

I did a google on "linux load management" and got a number of good hits. Give that at try and pay special attention to load balancing software. Here is one *random* hit that looks relevant to your problem:

http://lcic.org/load_balancing.html

jt1 12-30-2009 06:23 PM

Thanks for the help!!

I was wondering then how do these large cloud organizations allow for users to scale their VMs to large sizes, such as Amazon and Rackspace?

Quakeboy02 12-30-2009 08:10 PM

I believe that they have large separate database servers. Then there's akamai, which is essentially a load balancer in that it redirects you to the server nearest to you geographically. And each geographical location would have a load balancer and a server farm.

From your point of view, you need a DBMS that has enough horsepower to power you. Then you need enough servers to manage the total load. Finally, you need a load balancing solution that will redirect incoming connections to individual servers. IOW, you would present one virtual address to the world for your website, but the LBS would redirect the traffic as it sees fit to a real server.


All times are GMT -5. The time now is 07:51 PM.