IBM MQ 作為業界主流的企業級消息中間件，被廣泛應用于銀行、金融、政府、制造、零售等各種行業。它的通道傳輸是實作跨系統、跨地域的消息通信主幹的基礎技術，與 之相關的業務場景包括分支機構與總部的互聯、開放平台(Linux/Unix/Windows)與主機核心系統的上下行、雙活資料中心之間的資料同步等 等。
作者在支援大量MQ客戶案例的過程中，經常會遇到需要進行通道傳輸性能調優的情況。一方面，我們需要采用架構性、全局性的思維，從應用程式設計和代碼實作、硬體資源消耗情況（如CPU, 記憶體，存儲等）、網絡參數設定和調優、MQ設定和調優等方面着手進行分析、測試和調優（這是一個更為深入的話題是以這裡不作展開）；另一方面，作者也在進行一些有趣的嘗試，例如利用現有的MQ功能實作通道傳輸能力的垂直可擴充性，以便應用開發者或系統管理者通過簡單的指令便可實作通道傳輸能力的實時調整。
下面是此方法的簡要介紹、測試結果、擴充和收縮場景的操作指令，供各位參考。
As a predominant enterprise messaging middleware, IBM MQ has been widely used in various industries like banking, finance, government, manufacturing, retail, etc. The channel transfer of MQ is the foundation of messaging backbones between distributed systems and cross geographies. Related scenarios include interconnections between branches and headquarters, inbound & outbound connection between distributed platforms (Linux/Unix/Windows) and mainframe(z/OS), data replication between Active-Active data centers, etc.
I have been supporting a great number of MQ customer cases, and have often encountered situations in which MQ channel performance tuning is required. On one hand, we need to think architecturally and horizontally, perform analysis, testing and tuning from comprehensive angles like: Application design and implementation; Hardware resource consumptions (like CPU, memory, storage); Network configurations and tuning; MQ configuration and tuning. On the other hand, I have also done some interesting experiments, for example, using existing MQ functions to achieve the vertical scaling of channel throughput, so that either application developers and system administrators can using simple commands to dynamically adjust the channel transfer.
Below is a brief introduction to this approach, as well as the test result, commands for scaling up and down scenarios. For your reference.
