3 ms·
I am a customer of a competitive product, where we have a cluster of servers on port 25 that need load balancing. Pretty simple setup - if port 25 goes down for
by bks 10y ago
I am a customer of a competitive product, where we have a cluster of servers on port 25 that need load balancing. Pretty simple setup - if port 25 goes down for more than 2 minutes remove from DNS until it comes back online.
It's been incredibly stable and robust and a real life saver if you have gear in multiple geographic data centers that are not all with one provider. The competitors cost for the service is incredibly cheap hen you consider what it does - but the BIG GOTCHA is you need to use their DNS in order to use the load balancer.
And we move about 350 queries per second so the bill is pretty high (for our sized company). The place that I should be able to get an immediate impact and huge savings is that Cloudflare does DNS for "free".
I should be able to pay a modest but competitive rate for their traffic manager, but with no QPS it will be huge.
I just spoke to sales in the UK and they don’t actually have a good definition of the product yet. No pricing and really did not have their questions to ask prospects yet.
That being said we are already using them for pseudo-Traffic Management services we have a script tied to Pingdom.com that updates DNS from the port monitoring and adds and removes records in just a few seconds.
- kijin 10y agoIt's okay for port 25 to go down for 2 minutes because standards-compliant senders will retry when your server is back up. It's not okay for a widely used web service to go down for 2 minutes until you can update your DNS. Failover needs to happen within a few seconds, and most importantly, while the client is still connected to the load balancer. You can't do this using DNS with short TTL values.
- bks 10y agoI am extremely thankful for the robustness of SMTP.
- zzzcpan 10y ago> Failover needs to happen within a few seconds, and most importantly, while the client is still connected to the load balancer. You can't do this using DNS with short TTL values. No it doesn't and it's good enough. Humans are not stupid, they will try again, web browsers will try other IPs too. For custom clients it's not even a question, as they can implement transparent failover in the client itself. Essentially this is all about web browsers. The big problem with anycast routing is that you still have to rely on a single AS/company and have a huge SPOF, that will fail from time to time. A few hours of downtime every other year is as good as you can get, less than that would be pretty much impossible for such architecture. So, if you need anything better, you would have no choice but implement a DNS failover.