Question on IP addresses

[scientiam_interest]

Hi,

in the plans for the Cloud Service you mention IP Adresses and domains. Does this mean I will only be allowed to use it from one IP address and one domain if I take the basic plan?

Question is about development which takes place at another IP address then production server.

Is the domain restricted to a specific subdomain? Because in the staging release we would have a different domain then in the production domain.

Thanks for your help

[kamermans]

Hi,

The restriction of IP/domain is in reference to public-facing, production traffic, so you are free to test the service in development and staging environments. With regards to different domains, let's say you have the domain example.com; we treat subdomains like www.example.com and m.example.com and mail.example.com as the same domain, assuming they are still serving content for your company, not other companies. Similarly, alias domains like example.org that point to your main domain (and therefore serve the same content) are not counted as separate domains.

One more point of clarification on dynamic IPs. Some users have cloud hosting, shared servers, and other scenarios in which their public-facing IP address may change occasionally. If your account has a limit of one IP, then we shouldn't see traffic coming from more than one IP at the same time. When your IP changes, all the traffic will change from the old IP to the new IP in a very short time, so this still looks like one IP to us, and you are within the limits of your single-IP account.

[wurfl@limos.com]

How do you determine production vs staging/dev traffic? The IP restrictions will be no problem for us in production but we use many EC2 instances for staging and QA which is completely separate from production on another hosting service. On any given day we are likely to send traffic from more than 3 IPs, but all the non-production IPs will have a comparatively tiny number of requests. Is generating a separate API key and labeling it DEV helpful?

[kamermans]

Since your non-production servers will have a small number of requests coming from them, they will not be counted as production IPs, and in any case, we are looking for how many IPs are in use concurrently, so, for example, the number of IPs serving 90% of the traffic during a one-hour window. Also, we realize that there are some situations where there could be spikes of traffic from multiple IPs due to deployment and testing and things, so we take that into consideration. Using a dev key will help for cases where there is a manual review of the traffic.