📄 performance-faq.xtp
字号:
<document><header><product>resin</product><title>Performance FAQ</title><description></description></header><body><localtoc/><s1 title="How many concurrent users can a Resin server handle?" type="faq"><p>This is not a question that can be answered in a general way. Itis very dependent on the particular application that Resin is usedfor. Factors such as database usage, how the session object isused, the use of server side caching, and applicationarchitecture in general have a significant effect on thecapabilities of a website.</p><p>The best (and only practical) way to answer this question is toperform some benchmarking tests for your particular application ona server similar to the one that will host the website.The freely available httperf tool, as well as various others, areuseful for this purpose.</p><p>When using testing tools, 500 "concurrent threads" does not meanthe same thing as "500 concurrent users". A typical user is notconstantly making requests to the server. Typical usage involvesa request for a page (with possible subsequent requests forimages), and then a period of inactivity as the user reads orwatches the content that has been downloaded.</p><p>The ratio of number of users to number of threads again dependson the application involved. For example, it may be that theratio for an application is 50:1, meaning that 2500 users willuse at maximum 250 threads on the server.</p><p>Ideally, application benchmarks use "user scenario" scripts. Thescript imitates what a typical user wil do, including pausesbetween requests. This kind of script is useful for providing anaccurate picture of web server usage.</p><p>The primary configuration item in Resin for handling a greaterload is <a config-tag="thread-max"/>. The default in <code>resin.xml</code>can be adjusted upwards to handle increased load, the limit is determined bythe underlying operating system.</p><p>If anticipated load overruns a Resin server, either with CPUusage or with encountering OS thread limitations, clustering canbe used to add another server to share the load.</p></s1><s1 title="Does Resin use NIO for improved performance?" type="faq"><p>Resin does not use NIO - it uses JNI to handle low-level I/O calls with nativecode. The performance using this method was found to be much better than nio.</p></s1><s1 title="How does Resin use JNI?" type="faq"><p>The JNI code is compiled on the variousUnix systems when the <code>./configure; make; make install</code>step is performed during installation. Windows has precompiled dlls.</p><p>Resin uses JNI in certain critical performance areas, such as low level socketconnections and file operations. JNI is also used to interface with theOpenSSL libraries.</p><p>A significant benefit in particular is in Resin's ability to handlekeepalive's. With JNI, Resin does not need a thread for each keepaliveconnection. The low-level <code>poll()</code> (or <code>select()</code> ifpoll() is not available) functions are used. The end result is the possibility of many more keepalive's than if a thread was needed for each keepalive.</p><p>The fallback if JNI is not available is to use the JDK equivalents of thefaster JNI calls. Also, OpenSSL is only available through JNI.</p><p>Resin indicates that JNI is being used with a log message at startup:</p><example>Loaded Socket JNI library.</example><p>If JNI is not available, the log message is:</p><example>Socket JNI library is not available.</example></s1></body></document>⌨️ 快捷键说明
复制代码
Ctrl + C
搜索代码
Ctrl + F
全屏模式
F11
切换主题
Ctrl + Shift + D
显示快捷键
?
增大字号
Ctrl + =
减小字号
Ctrl + -