Load Balancing application servevr

[ShrutiTaywade]

Hi,

I would like to do an active active multi site deployment for Blue Prism 7.0 

• Number of application servers - 2 ( One for each site i.e. A and B site)
• Number of  Runtime Resources - 2 ( One for each site i.e. A and B site)
• Number of  Interactive client - 2 ( One for each site i.e. A and B site)

Image 1:

Image 2:

I have few queries:
1. I am planning to load balance these 2 application servers, which of the above will be correct architectural diagram.
​
2. If application server A goes down, will the runtime resource and Interactive client from the corresponding site would also go down? Because the unidirectional communication from Application Server A to its runtime resource (Image 2) will also go down. In this case, can runtime resource still listen to traffic from application server B?

3. Appreciate if you could elaborate how does the connection between application server to runtime resource work?

Thanks,
Shruti​​

------------------------------
Shruti Taywade
------------------------------

[Wing_LingLeung]

Hello Shruti,

I think the final answer really should come from your actual requirements. What you've currently laid out may not make sense at all in terms of your infrastructure and business workload requirements. There are considerations across data centers such as network latency that may make this type of failover to not satisfy your company's business needs. If you can, please enumerate what requirements you are trying to address and I am sure we (myself and the community) can support a deeper discussion.

With that said, I will point out a few things to your current questions.

1. I think the 2nd image is the more accurate one. While some LBed communication will go through the "pipe", There are still some direct instructional communication from the Application Server back to RRs. In image 1, there are no communication links from the App Server directly back to RRs not going through the LB.

2. The answer is no. Again, LBing this way may not serve your real requirements but in the instance of Image 2. If Application Server A goes down and there is something instructional that needs to be send to a RR, Application Server B will still be able to see, and can directly instruct, those RRs in Site A. Obviously the communication will be affected by latency across sites but the communication should be possible.

3. I think you already have some of the right references given the images you've posted. Here are guides that should help:
- Enterprise Infrastructure Guide: https://portal.blueprism.com/system/files/documents/v6.7%20Data%20Sheet%20-%20Infrastructure%20Overview%20Enterprise%20Edition.pdf
- Securing Network Connectivity guide, provides more detail on the communication between components: https://portal.blueprism.com/system/files/documents/v6.3%20Data%20Sheet%20-%20Securing%20Network%20Connectivity.pdf

------------------------------
Wing Ling Leung
Senior Product Consultant
Blue Prism Professional Services
------------------------------