Community Tip - Stay updated on what is happening on the PTC Community by subscribing to PTC Community Announcements. X
Hi Developers,
What are the HA features for KepserverEX ?
The current scenario is that we have defined channels, devices and data tags in a single Kepserver and its now connected to the Thingworx composer. Now we're looking for fail safe scenarios on the Kepserver machine. I couldn't find documents for this.
I came across link and learnt about Redundancy Master. I'm didn't quite get how to use this and this document seems to be outdated. We're currently using TWX_Industrial Connectivity 8.3.
Please suggest some useful guides.
Thanks and Regards,
Saran
Solved! Go to Solution.
Hi Saran,
The relationship between Kepware and ThingWorx Composer is NOT one-to-one; it is many ("N" number of) Kepware servers to one ThingWorx server. In other words, a ThingWorx composer environment has access to any Remote Things configured within the associated ThingWorx server, and each Remote Thing can represent data from diffrerent connected instances of Kepware servers.
I hope this helps clarify things.
In general NO, you would not want to use a load balancer between Kepware servers and ThingWorx servers because the connection between Kepware and ThingWorx is stateful and maintained. Load Balancers deal best with stateless HTTP requests because each request is more or less stateless, and so if the Load Balancer reroutes a particular request to a different server, there is no "connection setup" other than a TCP socket that must occur before the request can be answered. For multiple path / not single point of failure in the network between Kepware and Thingworx, better is to use NIC teaming on both the OS running Kepware and the OS running ThingWorx so that multiple paths of communications are possible but managed by the internal operating system. We can discuss more on the phone after you engage my team by emailing presales.support@kepware.com.
Thanks,
Sam
Another query is that should we use load balancer for the Kepserver if we're using two of them to account redundancy.
For kepware its one on one mapping to a TWX composer. So if the Leader composer fails how do the kepserver redirect to the Standby composer through a load balancer. Or is there some other mechanism.
In the document the architecture diagrams are showing Connection servers. Does that include Kepserver EX?
Regards,
Saran.
Hi Saran,
I manage Kepware's pre-sales engineering team. For these types of requests, please send an email to presales.support@kepware.com and a member of my team will assist you; often, we can often reply to you more quickly than if you were to post a message to this message board.
Regarding your questions, configuring Kepware and ThingWorx for high availability doesn't leverage simple, off the shelf ready-built tools. You'll need to work closely with your engineering contacts at ThingWorx and with my team or Tech Support in order to design an HA solution that meets your needs. The conversation will include topics such as your virtualization technologies (if any) and what those tools offer for OS protection, other third party tools that can help to guarantee uptime of a server like Windows Cluster, and then what features exist between Kepware server products and ThingWorx that can be leveraged to help protect data integrity; items like Kepware's Store and Forward feature and ThingWorx' ability to use Kepware RESTful Configuration API for enabling / disabling Kepware server connections.
So, it's a big discussion best had on the phone. We can use email to get the conversation started. Please follow-up with presales.support@kepware.com and we'll assist you.
Thanks,
Sam
Hi Saran,
The relationship between Kepware and ThingWorx Composer is NOT one-to-one; it is many ("N" number of) Kepware servers to one ThingWorx server. In other words, a ThingWorx composer environment has access to any Remote Things configured within the associated ThingWorx server, and each Remote Thing can represent data from diffrerent connected instances of Kepware servers.
I hope this helps clarify things.
In general NO, you would not want to use a load balancer between Kepware servers and ThingWorx servers because the connection between Kepware and ThingWorx is stateful and maintained. Load Balancers deal best with stateless HTTP requests because each request is more or less stateless, and so if the Load Balancer reroutes a particular request to a different server, there is no "connection setup" other than a TCP socket that must occur before the request can be answered. For multiple path / not single point of failure in the network between Kepware and Thingworx, better is to use NIC teaming on both the OS running Kepware and the OS running ThingWorx so that multiple paths of communications are possible but managed by the internal operating system. We can discuss more on the phone after you engage my team by emailing presales.support@kepware.com.
Thanks,
Sam
Hi @selsner,
Thanks for your quick and elaborate reply. It was very helpful for a novice like me in HA architecture.
I'll check on the topics you had mentioned with my team and once i get clear info on them i'll start the mail communication as soon as possible.
In the mean while is there any documents which can help me get started on the HA design for Kepserver.
Regards,
Saran