[Addition] Please read comments for most recent updates to this topic. [/Addition]
I host a friend’s personal WordPress site on Mosso, and it receives enough traffic to run some estimations. Overall the numbers look pretty safe. These numbers are only estimates based on the limited data I have. I will definitely publish additional data as I receive it.
First, let me point out that if your site is made up of static pages with no database connections, you shouldn’t have any compute cycle usage concerns. Roughly 20,000 unique visitors to www.surviveoutdoors.com = 100 compute cycles. It would take tons of traffic to go over your limit.
The personal WordPress site is averaging 100 compute cycles per 1733 unique visitors. So if your WordPress site receives 1733 unique visitors a day, over a month you will use up 3000 of the 10,000 compute cycle limit. A WordPress site that receives 5718 unique visitors a day will be right around the 10,000 compute cycle limit for the month. So how can I recommend Mosso, when we all know you could move a 5718 visitor/day WordPress site to a typical $10/month hosting account?
Why I still highly recommend Mosso
Mosso is still a great entry-level to mid-level hosting option. The support and communication has been wonderful. Being able to host both Windows and Linux sites through the same account is great. Mosso is the ideal option for someone who doesn’t feel comfortable having a managed VPS or dedicated server. Mosso is perfect for someone who wants an easy interface, to test out dozens to hundreds of small projects, moving the projects that take off to a different environment. While 5700 unique visitors a day to a WordPress site may seem like a low limit, it is plenty of room for general project startups and numerous small businesses.
However, when a low-bandwidth 5700 unique visitor/day WordPress site could simply pay $10/month somewhere for hosting, this does show that something is weird when it comes to the compute cycle method. Either that, or shared hosting companies take a loss on low bandwidth, high CPU usage sites.
Kyle, thanks for the post(s). We have WordPress sites that have tens of thousands of page views/day that don’t exceed the compute cycle quotas. It all depends on the plugins installed, if WP-Supercache is properly configured, etc. I also want to make it clear that this is our Cloud Sites offering – Cloud Servers, our on-demand computing platform starts at about $11/month. There are differences in the offerings that make Cloud Sites a different hosting solution:
Cloud Sites is built on a scalable, load-balanced, infrastructure – it is built to withstand the “Digg Affect”. A VPS won’t. So Scalability and Survivability are a huge part of this solution.
Cloud Sites is also a completely managed (and supported!) solution – you don’t mess with software patches – we do. You manage your application(s) and we handle the rest.
Cloud Sites also allows any Cloud Sites customer to instantly (nearly) become a reseller. We will even handle the billing for you, if you wish. Even end-user white-label support can be handled by us. Your choice.
And as you mentioned – the ability to use Windows and Linux on a single domain is a unique feature – and does not exist in the $10 price range.
We’re proud to offer a variety of hosting solutions so people can pick the one that best meets their value proposition – I’m glad you see the intrinsic value built into Cloud Sites (I was a Cloud Sites customer long before I became a Mosso employee!)
Thanks again,
Rob La Gesse
Director of Customer Development
Mosso | The Rackspace Cloud
210-845-4440
Thanks for the followup comment Rob.
The blog in reference has a simple quotation rotator, and the flexi pages plugin which simply changes how the navigation looks (http://srinig.com/wordpress/plugins/flexi-pages/). Akismet anti-spam plugin is not even turned on.
Are you advising that wp-supercache should be used? I wasn’t clear there.
Thanks again for keeping on top of this.
WP-Supercache will save a typical WP blog 80-90% of the database requests – which is a huge compute cycle saver. You can find our KB article on it here: http://help.mosso.com/article.php?id=328
Thanks again for choosing Mosso! And for sharing your experiences.
Rob
@Rob: I’ve been looking at Kyle’s numbers trying to determine if I can replace my server w/Mosso… and it’s a mix of WP, vBulletin and Drupal sites for the most part. You should change your calculator on the site to allow ‘hits per day/month’ or something… I was ready to go with mosso but the whole “compute cycle” thing is sketching me out!!
@Rob: I agree with sullo, I’m currently building a web application that is using WP as a backend alongside my own scripted CMS to handle a complex database with hundreds of users needing to store hundreds of entries each with 50+ variables including images (the CDN provider really got my attention as a major plus) . The information for all of these entries will be called in by other applications also hosted on the cloud account hundreds of times a day by each user. Thats ALOT of “computer cycles” needed in my opinion. And i understand that the scalabilty is a major point pushed about the cloud hosting but i’m not sure how many computer cycles i would need in the future and if the amount i need would send me way over budget. I plan to use WP-Supercache but i’m not sure if my tiny tiny dev team can handle the extra work load of modifying it to work with my CMS with the insane deadline were trying to reach anyway. Also, my team doesn’t have any hardware experience and neither does the client we are programming this for so the fact that all of the server side stuff is scaled for us is a major plus. My point is that while the plan looks like a GREAT deal up front, i get the impression that the tables will turn once my application goes into full force.
Thanks for your time,
John Moore
I will just add this on caveat to anyone considering Mosso that comes across this. The compute cycle on Mosso is not transparent at all so you don’t really know what’s going on. That’s the first problem. Worse is the lack of daily stats and the updating of the compute cycles. It is not what I would call timely. I have seen sudden jumps in the computer cycle in the UI from one day to the next (i.e. end of Monday it reads 500 and then end of Tuesday 5,000), yet none of the stats on the website suggest traffic spikes. This leads me to believe the calculation and updating is either lagged or not trust worthy. Since there is a lack of transparency in the calculation I have to say its the second.
Let’s also not forget that the number will often read “As of your last billing date X” and when you are on day Z (two days into the new billing cycle) it has all the compute cycles from the last month + what has just happened recently not allowing you to know what is going on.