ToC index

VPS2

Rad home

Forums

Blog

Virtual Private Server VPS Page 1

ToC index

VPS2

Rad home

Forums

Blog

Virtual Private Server VPS Page 1

Intro to Virtual Private Servers

NOTE: This guide remains a work-in-progress. It is based on personal experience, acquired from moving the site to a Virtual Private Server (upgrade from Shared webhosting).

So it's more like a Buyer's Guide to VPS .. than an Administrator's guide (which would resemble a Guide to Linux).

You won't find another guide like this. At least, I didn't. And I searched long-n-hard. (I wouldn't bother writing this if there were already something available.)

I learned much in the process (of upgrading to VPS), and plan to include everything here I wish I would've known when my research began. I'll continue to work on it a little every day .. 'til it's done.

Last updated » 11.May.2008 (began 07.April.2008)

I spent much time researching Virtual Private Servers .. after our previous web host (Lunarpages) booted us off their "Production server," and exiled us (banished us) to their "Stabilization server" .. for using "excessive server resources."

I also spent much time & effort trying (in vain, unfortunately) to reduce our use of server resources, so we wouldn't have to upgrade to a VPS. "Why didn't I want to move the site to a VPS?" you ask. Good question.

This brings us to the biggest (and only) negative associated with upgrading to a VPS » CO$T. A Virtual Private Server typically costs 5-10 times more than your average "Shared" web hosting account.

For example, we were paying $95/year with Lunarpages. Whereas the cheapest VPS plan at WiredTree (where we host now) costs $49/month. This plan is based on a memory/RAM allowance (limit) of 256-MB. (Unfortunately, Radified uses more than 256-MB.)

So you can see, at nearly $600 per year, the cost of a VPS represents a 600% increase .. over what we were previously paying for Shared web hosting. Fortunately I was able to find a coupon that provides a 10% discount, lowering our monthly fees to $44.

But that still costs times more than we were paying with Lunarpages. In other words, a year's worth of Shared-hosting buys only 2 months at a VPS. Big difference, huh? And no plans exist to bridge the sizable price-gap.

Note that an unManged VPS (which I discuss on a later page » Managed vs unManaged VPS Web Hosting) costs ~half as much as a Managed server.

If you have the mojo to administrate your own VPS however, you probably wouldn't bother reading this guide. And even if you *do* possess these skills, the benefits you get with a Managed VPS (such as » regular software updates, daily back-ups, and 24x7 monitoring) is still probably worth the extra cost.

If cost however, is NOT an issue for you (and you enjoy indulging your technolust), then there's no good reason NOT to upgrade to a VPS, because they are superior to Shared web hosting in every other way.

Even if your site does not actually *need* the extra horsepower offered by a VPS, your pages will still load faster with a Virtual Private Server .. due to allocation of dedicated server resources. And everybody knows fast-loading pages is one of the most desirable features a web site can have. (Slow-loading pages suk.)

Consider for example, the following statement from the WiredTree site (listed under the heading » Enterprise Hardware):

WiredTree strictly limits the number of accounts to ensure all environments are highly responsive.

Now a greedy web host can "over-sell" a VPS as easily it can a Shared web server [especially with Virtuozzo (virtualization software) .. harder to do with Xen, which works differently].

But my observation at our new VPS — thus far, anyway — has been noticeably faster-loading web pages .. especially those generated by a script .. such as the ones you find at the forum (which uses a Perl-based script).

Scripts tend to require ample CPU clock-cycles in order to generate web pages. The VPS which WiredTree uses comes with dual Xeon Clovertown processors (4 CPUs each) .. for a total of 8 CPUs per physical server. We used to have problems with CPU usage (back in the days of Shared web hosting). But not any more (at our new VPS).

The blog we use (Movable Type Open Source) serves static web pages, tho a script (also Perl-based) is used to create the pages, and this aspect of the blog is definitely faster (~3 times faster). No more waiting around for the pages to 'rebuild'. This is probably the single most dramatic speed-improvement I've noticed (creating blog entries with Movable Type).

Regarding Shared web hosting, this webpage (titled What is VPS Hosting?) claims Shared hosting providers cram as many as 500 customers onto a single server, while VPS limits the number "at most" to 20. I think this is an exaggeration, but you get the gist » many accounts are assigned to Shared web hosting servers, while relatively few get assigned to each VPS. (In the world of Web Hosting, less is more.)

Our old web host (Lunarpages, Shared webhosting) limited our use of server resources (both memory & CPU) to 1%. This suggests they put at least 100 customers on each server, since few sites will max-out and use the whole 1%, the entire time (24 hours/day).

Realize, too, that a single high-traffic site can overload a server faster than 500 low-traffic sites. So there's more to consider than merely comparing the number of accounts/sites assigned to each physical server. But in general, less accounts are better (for you & me, not our hosting provider), and dedicated resources definitely help prevent one busy site from bogging down our server.

Note that, with a Virtuozzo-based Virtual Private Server, tech support people sometimes refer to your individual VPS as a 'container,' while the physical server upon which your VPS runs is sometimes referred to as the 'hardware node.' So you can see how use of the generic term 'server' can be confusing.

In later pages, I'll discuss factors which can affect the responsiveness of our VPS, such as Server Load, CPU latency, and Disk Input/Output, which is typically the biggest bottleneck in virtualization. Some of these metrics are available only to our hosting provider, which can make troubleshooting responsiveness problems difficult (for us). But first, let's nail down the basics.

On the next page, I'll begin with the end, and share the decision I made regarding the selection of a VPS web hosting provider, along with my reasons for this decision. Those with time-constraints should find this info helpful, tho you should be a good capitalist and do your own research, as things in the web hosting world can change quickly.

NEXT » Selecting a Virtual Private Server Provider: VPS Web Hosting Decisions

For more along these lines, here's a Google search preconfigured for the query » virtual private server vps web hosting provider

This page was created with Dreamweaver, using XHTML 1.0 standards.

Valid XHTML 1.0 Transitional