Good question. It depends on the caching strategy.
As we essentially serve Gists we have two types of URLs one something like [gist-id]/ meaning the latest version of that particular gist and another URL is [gist-id]/[revision-id] which is a particular revision. For 'latest' URL we just do a redirect to particular latest version.
For first type of resource we have 60 seconds expire, however it is also not too expensive to re-generate it. For second type of resource (with particular revision) we have much higher expire time (several days).
By this type of caching strategy I assume we will get only 20% of all requests to our servers so saving is significant.
Since there is a 1005ms latency for a cache-miss, for the average user the loading times are shorter?
(based on your data on hit/miss ratio)
I was thinking to use Cloudfront to lower the latency (not costs) for dynamic resources, because in Italy there isn't an ec2 datacenter, but there is an edge location for Cloudfront.
No 1005 ms was not a cache miss it was a request to the origin server. Latency was 799 ms.
I assume CloudFront will contribute some latency, however in our case (we are using Heroku) our servers were already on Amazon platform, therefore CloudFront --> our server latency should be significantly lower. So CloudFront cache miss latency for the enduser will be not significantly higher than origin server latency for the enduser.
Thanks,
I've asked because I've read somewhere that an average request for a "Miss from cloudfront" object was ~1 second, and I thought that your data was confirming that value.
In terms of raw hardware, CloudFront is never cheaper than scaling your own servers. But when you compare the cost of CloudFront to running your own globally distributed network of servers, pushing media to edge locations, etc. it works out to be quite cost-effective.