As I am sure everyone is well aware at this point, Filmtracks is not monetizing its traffic effectively enough to sustain itself going forward. Since 2003, the site has required tremendous processing capability to serve the forums and review template technologies that were enhanced in 2006 and 2009, respectively, to demand even greater server resources. The loss of advertisement revenue (which came for over decade from companies that are either out of business now or struggling to some degree) forced the site to rely upon user donations that ultimately did not amount to enough consistent replacement income to suffice.

The bulk of Filmtracks' operating costs is owed to its server, a relative beast of a box that has, especially since 2010, admirably handled the processing demand of the advanced data systems I wrote in 2009 to run the back end of the site. Filmtracks can no longer afford this kind dedicated server. Ten years ago, such a dedicated solution of over $500 a month was not an issue. In fact, it was a necessity for projected growth. While traffic at Filmtracks has declined over the past few years, that loss of viewership had not correlated directly in a 1:1 ratio of decreased server load. Regardless of this reality, the site is, despite its advanced infrastructure (don't let the retro front-end fool you), not the kind of client that can demand a dedicated server any longer.

At the end of this month, I will be migrating the site to cloud-based hosting, which is essentially the glorified virtual/shared hosting of yesteryear with fewer costs to the service providers and scalability for the clients. I don't like cloud hosting, because the components that make up clouds, while existing in great quantities within the cloud, are individually slower. For instance, my dedicated servers have always had 10K-15K SCSI/SAS drives for improved read times on the site's data; conversely, most clouds run on inferior 7200 SATA drives. Likewise, whereas I had two 3.16 GHz processors running the site on the dedicated box, those in the cloud could very well be 1.5 GHz, and the theoretically equal provisioning four of them to your own allocation doesn't compare in real world application use.

Thus, the disadvantage of this move for Filmtracks is the potential for significantly decreased speed in loading, to the point where use of the site is cumbersome. The architecture of the forums in particular requires high CPU/RAM capacity, and without that option going forward, I will have to reevaluate the future of these areas. Hosting them off-site is not an option for security and legal reasons. I'm going to overload my allocated resources at the start and see how bad it is. I may be able to reconstruct the data to utilize more processor cores rather than demanding faster individual CPUs. It's impossible to say what will happen until the switch is thrown, because Filmtracks hasn't been on anything other than a dedicated server since March of 2000.

The advantage of a cloud environment is, obviously, cost and scalability. This move could save the site $250 to $300 per month. Cloud services also allow you to allocate more CPU/RAM to your solution on the fly as needed, which is (again, theoretically) a good option until you stop to consider that the CPUs and drives are inferior to begin with. There are "enterprise" cloud options which offer superior components. That would be like having a Borg collective consisting of assimilated nuclear physicists rather than a Borg collective consisting of assimilated dumbasses from a small-town Wal-Mart. Unfortunately, I can't afford an enterprise-level cloud solution, so Wal-Mart level it is for Filmtracks.

I have already provisioned the cloud server and installed Red Hat Enterprise Linux 6 as the operating system. Configuring a cloud server, I've discovered, is much more difficult than a dedicated one, because the latter always has a range of software packages already installed upon them by default as a courtesy. This cloud box had NOTHING on it other than the basic RHEL6 build... not even the damn utility with which to change the server's time zone. Very frustrating. Within a week, the software and configs will be completed and I can start transferring over the 400,000+ files. My goal is to complete the migration by the first week of May.

Those of you who have been through several of these server migrations with me (the last one was in May, 2010) know the routine. The forum shuts down for a day or so during actual transition. There is a 24 to 48-hour period during which traffic will be split between the two servers and no new content will be uploaded. I'll explain more on that later. I have no reviews to publish at this point anyway, other than three that still remain from prior to when I lost my last job in February of 2012. I have no idea when I will resume my writing schedule. These server migrations are always a challenge to execute well and they take up all my energy for upwards of a month each time. In this case, I'm also just started a new day job where the server situation is even more problematic (Windows 2008 Server/Microsoft IIS platforms are only used by 20%-25% of the websites in the world for a good reason).

I extend my thanks again to the readership that has stuck with the site these last few difficult years. And, of course, I remain very grateful for the assistance in donations. I am optimistic that I can manipulate the beneficial aspects of cloud hosting to create a good solution for Filmtracks going forward, time permitting. Regardless of whatever negative side-effects result from this downgrade in server, however, please understand that I no longer had any choice if my I wanted to keep the site going in a form even remotely resembling its current configuration.

Christian