There is a lot of speculation regarding how Google evaluates links and determines relationships between a linking site and its target. If you are a web developer and your clients are hosted on one of your servers, I would not make any effort to hide the relationship.
The prevailing thought amongst SEOs is Google understands IPs and therefore its important to not only change the IP address, but to ensure the new IP uses a different C block. This concept is years old and was employed as a first-step measure when Google first tackled the issue of site owners who were manipulating links. I strongly believe Google evolved past the linking C-block idea a long time ago.
Google is a very intelligent and experienced company when it comes to evaluating manipulative links. Here are a few examples of how Google can still determine a relationship exists despite varying your C-block:
-
when all your clients are hosted on the same server, and they all provide links to your site, the relationship is pretty obvious.
-
Google can ignore the IP addresses entirely and examine the nameservers. Many site owners use "SEO hosting" where they have one server with various IPs. The site owner specifically requests their IPs to use various C-blocks thinking Google wont recognize they are all using the same nameserver.
-
there are numerous other ways Google can relate sites. A specific combination of software is one example. Perhaps all the sites you develop are Joomla 1.7 with a certain combination of extensions. Perhaps you install a specific custom-created script or widget.
The bottom line is footer links are the least valuable type of link. You can expect Google to recognize the link and offer it some minimal value. There is not likely any significant improvement to the link's value by varying the IP. If you desire more traffic I would suggest creating code to solve a problem no one else has tackled. You can then earn authentic links from a variety of sources and be far ahead of the game.