High-Performance Computing for KNX Data Analytics: Xeon 28 Core 2.2GHz in EcoStruxure Building Operation

[jekede3166]

Hello KNX Community,

We are designing a centralized management system for a large-scale building automation project utilizing a extensive KNX installation. Our goal is to use EcoStruxure Building Operation to not only control the system but also to perform long-term, data-intensive analytics on energy consumption, occupancy patterns, and predictive maintenance.

To handle the potential data load from thousands of sensors and the complex processing for analytics, we are considering a server with an Intel Xeon 28 Core 2.2GHz processor (with 20GT/s UPI) for the backend.

I would like to understand the community's perspective on the following:

1. Data Processing & Historical Logging: For those running large KNX installations with years of historical data, has anyone evaluated the need for such a high core-count processor? Is the primary bottleneck for large-scale analytics typically the EIB/KNX TP bus itself, the database I/O, or the computational power needed for processing the data?

2. Virtualization & Consolidation: We plan to host the Building Operation server, a dedicated historian database (like a SQL server), and potentially other building management services on a single physical host using virtualization. Would a Xeon with 28 cores provide significant headroom for consolidating these KNX-related workloads, or is it overkill for the primarily I/O-bound nature of KNX data acquisition?

3. EcoStruxure Integration: From a best practices standpoint, are there known benefits or potential compatibility considerations when pairing the EcoStruxure Building Operation platform with such a high-parallelism CPU architecture? Does the software efficiently leverage multiple cores for its internal processes and analytics engines?

4. Future-Proofing for KNX IP & Secure Connections: As KNX IP and cloud connectivity become more prevalent, having headroom for encryption, data translation, and secure session handling is crucial. Could the high core count of this Xeon be beneficial for offloading these tasks without impacting core control logic performance?

Any insights from your experiences with large deployments, performance benchmarking, or server sizing for data-heavy KNX projects would be greatly appreciated.

Thank you,