Punch Architecture Training Module¶
Abstract
This module provides insights for architects/integrators on standard architecture patterns, mechanisms and constraints. It helps make system andsolution design choices, sizing estimate and capacity management, and identify the solution evolution.
It covers High-Availability, Scaling/Load-Balancing, Data lifecycle management, Database engines specifics, Sizing/capacity management topics.
Before You Start¶
Welcome to the Punch Architecture 'training module'.
Let us be truthful from the start: there is no such thing as an universal architecture design method, nor a design checklist to make it an easy task to design a sound and optimized solution fo every use case. With or without the punch.
That is why this training only aims at improving the attendees understanding of punch framework components, and some usage patterns of their integration to solve typical constraints such as high-availability, data protection against failures, short-term vs long term data management, dual-site synchronization for site disaster handling.
This training is intended to help solution/system architects that integrates the punch in their design. It can also help solution/system integrators to better understand subtle technical capabilities and constraints of Elasticsearch, Kafka, Zookeeper, Spark, punchlines, Clickhouse etc..
With that you should take the best out of your punch and design an evolutive solution that will resist years of lifetime in production.
This training requires that you are familiar with the concepts, components and general configuration of the punch. Going through the High Level introduction Training module is a must.
Training tracks¶
-
Track 1: Specifics of main Punch platform framework components:
clustering/Data life-cycle/High-Availability/Replication, deployment constraints, Scalability for Zookeeper, Kafka, Elasticsearch
-
Track 2: Punch Design Patterns and Reference Architecture:
- The dual-site synchronization/Disaster management pattern for central sites
- The centralized monitoring architecture including remotely monitored collection sites
- The 3-nodes pattern for remote collection sites (LTRs)
-
Track 3: Architectural Design Issues Highlight:
- sizing key points
- Database engine considerations and integration with custom application/user interface development
- N-Tiering for (real?) security hardening