Running dozens/hundreds of websites can be an expensive business, hosting alone can cost upwards of $250 a year for each site. SEOs will often look to host each website on a different IP range or in a totally different country. You may want to host country-specific sites in their target country, but not want to have to replicate your CMS or custom platform on every hosting server. A slightly darker reason for spreading your sites across different hosts would be to manage micro-sites that appear independent and use them to point link juice into your website.

One experiment I made last year was a script to auto-upload and update WordPress blogs no matter where they were hosted. Most web hosts don’t offer SSH access though and managing a large number of files via a PHP FTP class can get very messy. This also didn’t reduce costs, as each hosting account still required a MySQL database and all the pre-requisites for running WP.

A better solution comes in the form of copying the way that web proxies work. You’ve probably come across web proxies when investigating duplicate/scraped content; they basically allow someone to view a web page without leaking their browser, location or IP details to the website owner or their analytics provider. They’re mostly used by school kids to circumvent website restrictions, the tin-foil hat brigade, and by those living in countries with repressive regimes. The part that we’re interested in though is how the script visits a URL, temporarily stores the HTML, images plus JS/CSS and then displays the page without ever leaving the proxy site. Any web page can be viewed on the proxy site’s domain, without the proxy user noticing any difference.

Using these principles, I’m going to show you how you can manage all of your websites on a single, good quality, hosting account, whilst making them appear to be hosted on different hosts.

Step 1
Find a really reliable web host with high uptime and powerful servers. I use VPS Link for my VPSs and LiquidWeb for my hosting accounts. The larger your planned network will be, the more you should invest in this single site hosting. This will be our “hub site”.

Step 2
Create a simple Content Management System that can load different templates and content based on the URL parameter parsed to it (e.g. ?site=1). Alternatively you can use an existing CMS and adapt it to load data from different databases, depending on the variable sent to it. For WordPress you can make the following change to wp-config.php:

define(’DB_NAME’, ‘database’); // The name of the database
define(’DB_USER’, ‘username’); // Your MySQL username
define(’DB_PASSWORD’, ‘password’); // …and password
define(’DB_HOST’, ‘localhost’); // 99% chance you won’t need to change this value

To:

switch ($_GET[site]) {
case “1″:
define(’DB_NAME’, ‘database1′); // The name of the database
define(’DB_USER’, ‘username1′); // Your MySQL username
define(’DB_PASSWORD’, ‘password1′); // …and password
define(’DB_HOST’, ‘localhost’); // 99% chance you won’t need to change this value
break;
case “2″:
define(’DB_NAME’, ‘database2′); // The name of the database
define(’DB_USER’, ‘username2′); // Your MySQL username
define(’DB_PASSWORD’, ‘password2′); // …and password
define(’DB_HOST’, ‘localhost’); // 99% chance you won’t need to change this value
break;
case “3″:
define(’DB_NAME’, ‘database3′); // The name of the database
define(’DB_USER’, ‘username3′); // Your MySQL username
define(’DB_PASSWORD’, ‘password3′); // …and password
define(’DB_HOST’, ‘localhost’); // 99% chance you won’t need to change this value
break;
default:
define(’DB_NAME’, ‘database’); // The name of the database
define(’DB_USER’, ‘username’); // Your MySQL username
define(’DB_PASSWORD’, ‘password’); // …and password
define(’DB_HOST’, ‘localhost’); // 99% chance you won’t need to change this value
break;
}

I’ll be releasing a free Open Source (GPL) application shortly which will manage this and your blogs for you using a simple browser based GUI.

Step 3
Register as many free or cheap hosting accounts as you’d like, each will be displaying a unique website. You’d be surprised how many companies offer free hosting in exchange for a simple advertisement or link on your site. Each host must support PHP and use Apache Web Server (this qualifies 99% of web hosts). No MySQL databases or special PHP functions are needed. These will host our “node sites”.

Step 4
Now this is where the magic happens. We’re going to upload just 2 files to each of our new node site hosting accounts. After this, you’ll never have to FTP into or touch the hosting account again thanks to our proxy technique. The files that we’ll be uploading are a .htaccess file and a PHP script. I’ve reduced this technique down to just needing a .htaccess file in the past, although this requires a mod_proxy plugin which most free or cheap hosting companies will not support (Hat tip to Mike for pointing this out). Anyways…..

Create a simple text file in Notepad or equivalent named .htaccess and paste in the following:

Options +FollowSymLinks
RewriteEngine On
RewriteBase /
RewriteCond %{REQUEST_URI} !^/blog-app.php
RewriteRule ^(.*)$ blog-app.php?secret=monkey&u=$1 [PT,L,NE,QSA]

Once uploaded, this file will activate the mod_rewrite module in Apache (it’s the module that handles your 301 redirects usually) and points any requests (other than for /blog-app.php) to /blog-app.php?secret=monkey&u=$1 without changing the URL in the user’s browser. You’ll notice 2 variables in the URL we’re pointing to, the first (secret) is a security key that I’d recommend changing to a unique alphanumeric password and the second replaces $1 with the URL path requested. e.g. http://www.example.com/dir/test.html gets internally pointed to http://www.example.com/blog-app.php?secret=monkey&u=dir/test.html.

Step 5
Now the next bit of code is a big’en! It’s work-in-progress and therefore a little messy, although basically acts as our proxy. After checking that the request is genuine (using the secret variable), it will proceed to grab content from our hub site. So http://www.example.com/dir/test.html goes to http://www.example.com/blog-app.php?secret=monkey&u=dir/test.html which requests content, images and files from http://www.hubsite.com/dir/test.html?site=2.

Create a new plain text file called blog-app.php and paste in…

< ?php
error_reporting(0);

// BEGIN VARIABLES
$secret = "monkey"; // Should be the same secret key as in the .htaccess
$proxyto = "http://www.hubsite.com/"; // Where to proxy to
$proxyfrom = "http://www.example.com/"; // Where to proxy from
$siteid = "2"; // This Website's ID
// END VARIABLES

if ($_GET[secret] == $secret)
{
$q[] = "site=".$siteid;
foreach($_GET as $name => $value)
{
if ($name != ’secret’ && $name != ‘u’ && $name != ’site’)
{
$q[] = urlencode($name).”=”.urlencode($value);
}
}
if (is_array($q)) { $q = “?”.implode(”&”,$q); }
$u = $_GET[u];
if (strpos($u,”.jpg”) || strpos($u,”.jpeg”) || strpos($u,”.gif”) || strpos($u,”.png”) || strpos($u,”.bmp”) || strpos($u,”.swf”)) {
$q = FALSE;
$u = str_replace(” “, “%20″, $u);
}
$url = $proxyto.$u.$q;
$server = $proxyfrom;

class WebPage {

var $URL;
var $pageData;
var $headers;
var $static;
var $currentServer;
var $scriptName;
var $varName;
var $updatedPageData;
var $relDir;
var $fp;

function WebPage($URL, $static, $currentServer, $scriptName, $varName, $relDir){
$this->URL = $URL;
$this->currentServer = $currentServer;
$this->static = $static;
$this->relDir = $relDir;
$this->pageData = “”;
$this->varName = $varName;
$this->scriptName = $scriptName;
}
function openLink(){
if ($this->fp = fopen($this->URL, “rb”)) {
$fileheaders = stream_get_meta_data($this->fp);
$this->headers = $fileheaders;
}
else {
$error404 = TRUE;
}
}
function getRawPageData(){
fpassthru($this->fp);
fclose($this->fp);
return;
}
function getPageData(){
global $proxyto;
global $proxyfrom;
global $css;
global $u;
$updatedpd = str_replace($proxyto,$proxyfrom,$this->updatedPageData);
$strpath = substr($u, 0, strrpos($u, “/”));
$noslash = substr($proxyto, 0, -1);
$strpath = substr($strpath, 0, strrpos($strpath, “/”)).”/”;
$updatedpd = str_replace($proxyfrom.”../”,”../”,$updatedpd);
$updatedpd = str_replace(”url(’../”,”url(’”.$proxyfrom.$strpath,$updatedpd);
$updatedpd = str_replace(”url(../”,”url(”.$proxyfrom.$strpath,$updatedpd);
$updatedpd = str_replace(”url(’/”,”url(’”.$proxyfrom,$updatedpd);
$updatedpd = str_replace($proxyfrom.$proxyfrom,$proxyfrom,$updatedpd);
$updatedpd = str_replace($noslash.”\”",$proxyfrom.”\”",$updatedpd);
$updatedpd = str_replace($noslash.”‘”,$proxyfrom.”‘”,$updatedpd);
return $updatedpd;
}
function processPageData(){
$this->pageData = “”;
while( !feof( $this->fp)){
$this->pageData .= fgets($this->fp, 4096);
}
fclose($this->fp);
$delim[]=’”‘;
$delim[]=”‘”;
$delim[]=”";
$pre[]=”src=”;
$pre[]=”background=”;
$pre[]=”href=”;
$pre[]=”url\(”;
$pre[]=”codebase=”;
$pre[]=”url=”;
$pre[]=”archive=”;
$this->redirect($pre,$delim);
}
function fileName(){
return eregi_replace(”[#?].*$”, “”,
eregi_replace(”^.*/”, “”, $this->URL));
}
function redirect($prefixArray,$delimArray){
$a = $this->pageData;
$name = $this->currentServer;
$fileDir = $this->relDir;
foreach($prefixArray as $prefix){
$start2 = stripslashes($prefix);
$start = $prefix . “[ ]*”;
foreach($delimArray as $delim){
if(eregi($prefix . “[ ]*” . $delim, $a) && ($delim == “” ? eregi($prefix . “[ ]*” . “[a-z,A-Z,/,0-9]“, $a) : 1)){
$a = eregi_replace($start . $delim . “//”,
$start2 . ‘\a’ . “//”,
$a);
$a = eregi_replace($start . $delim . “/”,
$start2 . $delim . $name,
$a);
$a = eregi_replace($start . $delim . “http://”,
$start2 . ‘\a’ . “http://”,
$a);
$a = eregi_replace($start . $delim . “mailto:”,
$start2 . ‘\a’ . “mailto:”,
$a);
$a = eregi_replace($start . $delim . “#”,
$start2 . ‘\a’ . “#”,
$a);
$a = eregi_replace($start . $delim . “javascript:”,
$start2 . ‘\a’ . “javascript:”,
$a);
$a = eregi_replace($start . $delim,
$start2 . $delim . $name,
$a);
$a = eregi_replace($start . ‘[\]a’,
$start2 . $delim,
$a);
}
}
}
$this->updatedPageData = str_replace(”\a”,”",$a);
}

}

function processHeaders($headers, $filename, $mime_dl, &$type, &$isDown, &$isHTML, &$isImage){

global $url;
$openfp = fopen($url, “rb”);
$headers = stream_get_meta_data($openfp);

foreach ($headers as $header) {
foreach ($header as $headervalue) {
if (strpos(strtolower($headervalue), “image”)) {
$isImage = true;
}
if (strpos(strtolower($headervalue), “text”)) {
$isHTML = true;
}
if (strpos(strtolower($headervalue), “css”)) {
$headers[] = “Content-Type: text/css”;
$isHTML = true;
$css = 1;
}
}
}
return $headers;
}

$page = new WebPage($url,true,$server,”blog-app.php”,”u”,$relDir);
$page->openLink();
$head = processHeaders($headers,$file,$mime_dl,$type,$isDown,$isHtml,$isImage);
foreach($head as $h) header($h);

if($isHtml){
$page->processPageData();
echo $page->getPageData();

}else{
if($isImage)
$page->getRawPageData();
}

}
else
{
$error404 = TRUE;
}

if ($error404) {
header(”HTTP/1.0 404 Not Found”);
print “< !DOCTYPE HTML PUBLIC \"-//IETF//DTD HTML 2.0//EN\">

Not Found

The requested URL “.$_SERVER[PHP_SELF].” was not found on this server.

“;
exit;
}
?>

(Thanks to the PHP Proxy project for some of this code)

Step 6
Edit the above files as needed and then upload across your node site hosting accounts. You’ll only need to change the $proxyfrom and $siteid variables for each site, where $siteid is a unique ID and $proxyfrom is the root URL of the domain used on that specific node site.

Hopefully this will be of some help to the advanced readers of my blog. If this might as well be Egyptian hieroglyphs to you, keep an eye out for the release of my Blog-App PHP application that puts all of this behind a simple user interface.