The platform is experiencing some issues - we are very sorry about it. Our engineers are already working on stabilizing the system. If you would like to get live updates on the status please subscribe on our status page here.
DNS Load Balancing
Created on 2020-10-05 14:44:04; Last updated on 2021-10-17 00:00:45
Hosting and redundancy
Babelway infrastructure is hosted externally. Babelway has agreements with 2 hosting providers:
- Combell, a recognized Belgian hosting company. Combell uses several physical premises in Belgium. Premises have been audited by an independent consultant mandated by Babelway.
- Amazon, a recognized International company and cloud inventor. Babelway subscribes to the Amazon Web Service (AWS) allowing Babelway access to virtual servers "on-demand". The servers are physically located in Ireland and other locations around the world.
Babelway has contracted reliable partners on strict terms and has installed redundancy between its 2 data centers to maximize availability and reliability. In the event of downtime of one of the 2 data centers, Babelway can switch all data traffic to the other data center. Current limitation of the redundancy are:
- SelfService application (human access to user hubs) are only active on one data center. In case of unavailability of this infrastructure, messaging services can continue but human tracking or maintenance is not available. A manual process allows Babelway to switch the SelfService application to the other infrastructure.
- Gateways to external systems based on physical IP addressing would also be interrupted. We recommend that users use a URL locator instead of IP addressing wherever possible.
DNS load-balancing and fail-over
Babelway always maintains the configuration of the 2 data centers in sync. All customer configurations are deployed concurrently on both infrastructures. The load balancing and the fail-over between data centers are performed at the DNS level. The messaging engines deployed in the 2 data centers have the ability to work in complete autonomy in both active / active or active / passive modes. This mechanism is used by Babelway for fail-over as well as scalability needs.
To fully leverage the Babelway fail-over mechanism, traffic with Babelway gateways is based on DNS (logical addressing) and accepts traffic with each of the public IP address published by Babelway.
For outbound traffic, Babelway guarantees that all traffic will be issued from one of its public IP addresses.
For inbound traffic, Babelway guarantees that all DNS entries for public services, like ftp.babelway.net, ws.babelway.net or as2.babelway.net will be resolved to one or several of its public IP addresses. At any time, DNS will be resolved to 2, 3 or 4 servers. A client's software connecting to Babelway should be able to fail-over to the next IP should a server be unreachable (like all browsers do). Babelway reserves the right to change these allocations.
Taulia special infrastructure: Taulia has a very specific setup, where from the beginning Babelway decided to have for each region (US and EU) 2 full infrastructure, one active and one passive (8 servers = 2 regions * 2 actives and 2 passives).The 2 passives were never activated but the ip are reserved and the configuration is maintained.
0 people found this helpful.