Unfortunate typo in title but I can vouch for scalr.
Used it since the early paid beta. Compared to the likes of rightscale it's a bargain. Alongside all the expected things like ami creation and cluster (aka farm) configuration the scripting features and automatic DNS management make life much easier. Their quality control used to be a bit hit and miss and it sometime felt like we were testing their functionality for them. I also had the impression their European DNS performance was poor (never validated). For the money though they were still a cost effective way of managing large ec2 clusters.
The open source version used to be a real pain to setup. Given the relative cost of setting up and running your own install I'd go with the paid version.
Have never written a whole book but back in the day I contributed a number of chapters to various Wrox books on ADO.NET. Glad I did it but the Wrox editorial process was a bit of a shambles and it showed up in many of their books at the time.
It seems a bit pricey compared to other options but they handle things like charge backs and locale specific tax. Whether that justifies the higher commission rate is debatable.
Some things I didn't like about it:
(1) For new / low volume users there's a mandatory telephone number field.
(2) Once the transaction is complete it's not possible to automatically redirect back to your site.
Have you tried geonames? I've not specifically used it for the US but that kind of information is in there for other countries. Only drawback of using geonames is that it can be a bit of effort getting what you want out of their feed.
Having used OpenX for several years I wouldn't recommend it.
We used both the (free) hosted version and also hosted it ourselves. Either way it had a habit of hanging for no apparent reason. This didn't seem related to site load though, sometimes it just dozed off. When it was running "well" performance was OK but never earth shattering. To answer your specific question; with the right hardware 40k views shouldn't strain OpenX though it depends on the number of ads you have per page and whether you can use optimisations like Single Page Call. To give you an idea we used a single medium instance on EC2 and it worked fine, even coping with some meaty traffic spikes during TV advertising. We averaged 3 ad units per page.
Putting aside performance issues it was my experience that OpenX wasn't that reliable a delivery platform. It would frequently under deliver for no apparent reason.
Delivery issues are well documented if you search for "openx under delivering". It's been going on years. Thinking we were doing something stupid we got in one of the leading OpenX consultants who confirmed it's a "quirk" of OpenX. The way to "fix" it was through fudging the system with bizarre combinations of remnant ads and weightings. We upgraded several times to a version that was slated to fix the problems but it never did. The last time we upgraded was early this year (can't remember the specific version) but that didn't help.
If all you're doing is using your own Google Ads (or other in-house ads) then under delivery is less of an issue. If you're selling inventory to 3rd party agencies you'll soon become fed up with explaining why you continuously under deliver. Not only is it embarrassing it's bad commercially as well.
I don't have a great deal of experience with other ad platforms but if I ever need to do ads again I'll look into Google DoubleClick for Publishers (DFP). It's free upto a large number of monthly impressions and based on a comparison I did at the start of this year it can do pretty much anything OpenX could do - the only slight headache was targeting in DFP is not as straight forward as OpenX. Other than that I'd happily have migrated to DFP but just didn't have the time.
Used it since the early paid beta. Compared to the likes of rightscale it's a bargain. Alongside all the expected things like ami creation and cluster (aka farm) configuration the scripting features and automatic DNS management make life much easier. Their quality control used to be a bit hit and miss and it sometime felt like we were testing their functionality for them. I also had the impression their European DNS performance was poor (never validated). For the money though they were still a cost effective way of managing large ec2 clusters.
The open source version used to be a real pain to setup. Given the relative cost of setting up and running your own install I'd go with the paid version.