Cloud Servers

M

MustangV10

Guest
What's your opinion on cloud servers? I think they're pretty cool, and also pretty unique at this moment in time. I've only seen one or two company's that provide cloud hosting, Site5 and OVH...then again, I don't browse hundreds of hosts a day or deliberately look for cloud hosting.
 


This is very abstract to begin because of the varying definitions in what makes something "Cloud". Being a Project Manager & Software Engineer on a product that is currently being ported the "Cloud", I now have a better understanding.

It isn't the server itself that is "Cloud". It IS possible to have hosting in a Virtual fashion in the "Cloud" (take Linode for example), but what does that really mean? At the end of the day hardware is hardware, and a VPS is a VPS, right? In the instance of "Cloud" hosting, what this means is that you can dynamically grow your storage, memory, etc. as your needs grow. But wait! Haven't we already been doing this for years in current VPS solutions? Why, yes we have! When you're VPS needed to grow, your hosting provider would continue to allocate you more space. As all the VPS's continued to grow on a particular device/server, the host could then add more hardware to the server (memory and storage are about the only things), or he could port all the Virtual environments to a new server.

Okay, that's great, so how has the "Cloud" changed hosting? To begin you aren't limited to a single Virtual environment. Instead you are given an interface, or a portal, to control your hosting needs. This allows you to create as many Virtual environments as you want without the need of filling out a new order form. Simply go into your portal, set it up with your desired flavor, and go. Sure this does sound close to what we have always been doing, but there is a difference...billing. Before I begin, I must point out that a lot of the more popular "Cloud" providers have packages available for you. These packages include a set amount of storage, bandwidth, etc. along with a packaged monthly price. You have the ability to easily expand beyond this, and in fact you can without even having to request the expansion in many cases. "Cloud" has streamlined the billing process.

That's great, but how and why does it help me? Well, "Cloud" billing can be a double-edged sword. First you have full control of your Virtual environment creations. This is great as you don't have to fill out an order form each time, etc. You simply add what you need and your next billing period they just add the additional resources into your bill. You consistently have one bill, though the price does vary, and it is consistently on the same billing date (something that can not be said for standard hosting where each time you sign up for a new service you end up having multiple bills on different dates...typically, but there are cases that break this standard). The reason it is double edged, is this can also hurt you. In many cases a user pays for the amount of time the server is online/used (typically a small small amount per minute), you pay for each GB of space used, they will _SOMETIMES_ average out how much processing power you are using and charge for it, and they may typically charge for bandwidth consumed. For a small company who may be getting a lot of hits on the server, and a lot of storage consumption, this form of billing can be hurtful. However, you can easily see how this will allow you to grow with ease. No constant monitoring of your resources is needed here. You're hosting environment grows with you, and if you play the game well, you will actually come out ahead!

Now, if we are talking about a "Private Cloud", concepts change a bit. In a "Private Cloud" you are in control of the hardware, and the resources allocated. Typically this is pretty standard in the form of setting up a Dedicated server and allocating some Virtual environments to it. The difference being is that you create a system that allows those enriionvments to start small, and grow with you. After that all that is up to you is expanding the hardware as you need it. A "Private Cloud" really begins to gain definition when we talk about "Cloud" services/applications.

So that is all well and good, and we've covered "Cloud" hosting in the form of servers or Virtual environments. However, there is another way to define "Cloud", and that comes down to applications. We've seen this concept already with the SaaS (Software as a Service) approach. A business creates some kind of software, you buy it, and they throw it up on a domain to host it for you. In Software as as Service if there is something trackable in the system, such as a project management based SaaS. In a PM SaaS you might typically buy a package that says you can only have 3 projects, 2 users, and a SVN/GIT repo on each project. The "Cloud" based approach would allow you to only be charged for the projects and users you have, as you might not need 3 projects or 2 users, or you may need 3 users and only 2 projects.

What about the "Private Cloud" approach? In this instance, lets actually use the software I develop. I develop a Change and Configuration Management Database (CMDB) style software. This software allows for easy tracking of your I.T. Infrastructure on an ITIL platform (ITIL is a standard for internal tracking of hardware, people, software, etc...google it). Essentially, a company would create an entry in the software for each server, software, service, person, datacenter, cost center, etc. You put as much or as little in it as you want. You can also create relationships, so you know if you take this switch offline, that server goes down, which affects these services, and ultimately affects these customers. Short and simple, though it can get complicated. For each thing you put in the software (each server you track, each pserson, each service, essentially each entry) it increments by 1. We call each entry a CI. In the standard software you have a License that says you only get x amount of CIs and if you reach your limit you can not create more. You must contact support, billing will contact you, you pay for a new license and that gives you more CIs.

That is great, but there are some bumps here. First, you must know how many things in your organization you need to track. Give to low a count and you will need to buy additional CIs, which means you will go over your original budget. Count to high and you will have ended up paying more than you needed to. In the "Cloud" we license you out an amount of CIs based on your estimate. You can go low because once you reach your limit you just get an alert. You can continue creating more entries in the software, and your next billing cycle (6 months or 1 year) we run a report and charge you accordingly.

In a "Private Cloud" the same concept applies. You buy the hardware and we give you an Image to load up on the hardware. This image contains the software. At the end of each billing cycle we have you send us a report and we bill you accordingly. Fail to send the report, or fail to renew, and service ends. Though we can't take the hardware away since it is yours, and we can't force you to uninstall the software, you loose access to updates and you loose access to support. Already you see the difference. In a "Public Cloud" we would simply shut you down, but in a "Private Cloud" we just refuse to support you.

So that is all well and good, but surely there is something else that plays into this? Well, as a matter of fact there is! In a "Private Cloud" you actually would typically start out with TWO servers. One server would be fairly small. Wouldn't require much processing power or anything really. This server hosts the software front in. The portal you actually see. This smaller server is connected to other devices/servers on the network. These devices are where your processing power go. Each of the additional devices do processing, run jobs, store information in databases or flat files, etc. Your portal server takes information you give it and sends it to one of these devices. These devices store it, process it, run jobs against it, etc. When you are viewing stuff stored in the software what is really happening is the portal server is going out to these additional devices and getting the information you need. A lot plays into this but essentially each device has an ID, and there is a small local database that the portal server looks at and says, "Oh you need this information? Well, that is on this server. Let me get that for you." In the case of my software you can run jobs or processes. Most of these are automated and balanced out with all the devices (meaning on a rotation each device gets the next job), but there are some instances where eyou actually get to pick what device runs a job. At the end of the day it is a load-balancing type process. Need more space or processing power? Simply add a new device to the network, load up the software needed (not the portal software, but the agent software), and tell the portal where this new agent/device is located. Thus, expansion in an private environment.

There is of course other things that play into "Cloud", and everyone has their own opinion, but hopefully I've covered off on the basics here and you have a better understanding of how the could works. At the end of the day there isn't really a "Cloud" server, but more like a cluster of servers creating the "Cloud" environment. "Cloud" has helped streamline expansion and billing and if played correctly can save a company thousands of dollars each year on web and application hosting, as well as storage.
 
Did you actually type all of that? I'll be there for years reading all of that, lol.
 
Did you actually type all of that? I'll be there for years reading all of that, lol.

Haha, yes. If I'm going to help/answer someones question, I'm not going to go copying and pasting stuff unless it is quoting, a code example that I trust, etc. Because how do I know what I'm posting is actually accurate? Better for me to post what I know and hope it helps, than to post something by someone else and mislead :) Also, I talk a lot :p
 
Haha, yes. If I'm going to help/answer someones question, I'm not going to go copying and pasting stuff unless it is quoting, a code example that I trust, etc. Because how do I know what I'm posting is actually accurate? Better for me to post what I know and hope it helps, than to post something by someone else and mislead :) Also, I talk a lot :p

I see. Very interesting read when I had the time to look through it. Don't you feel a little worried about storing data online though?
 
Personally, I much prefer cloud servers. I currently have one situated in the UK and it's so much faster than any other server I've had. I like the way they're set out, where if one goes down, you're automatically put onto another server etc etc
 
I like cloud idea because it is easy to access from everywhere. Security is kept by security access and it suites me.

I heard about NailIt cloud for small business environment. Have anybody idea if it can run more than 20 users?
 
An example of an application that is cloud based would be gmail, hotmail, etc. Where your email application is actually running and stored in the cloud (Internet).This is typically used by companies to lower onsite upkeep costs and for business continuity in case of a disaster.
 

Members online


Latest posts

Top