java - 从服务器下载大文件需要很长时间

Question

我正在尝试从 Kaltura 下载文件(.mp4)，文件超过 100mb，下面的代码需要花费大量时间来下载。

还有其他方法可以改进吗？

try {
        URL url = new URL("https://www.kaltura.com/p/....");
        InputStream in = new BufferedInputStream(url.openStream());
        ByteArrayOutputStream out = new ByteArrayOutputStream();
        byte[] buf = new byte[8192];
        int n = 0;
        while (-1 != (n = in.read(buf)))
        {
            out.write(buf);
        }
        out.close();
        in.close();     
        res.setContentType("application/octet-stream");
        res.setHeader("Cache-Control", "no-cache");
        res.setHeader("Content-Disposition","attachment;filename=Smac_03_48.mp4 (iPad).mp4");
        res.setStatus(200);
        res.getOutputStream().write(out.toByteArray());
    } catch (Exception e) {
        e.printStackTrace();
    }

Answer 1

问题是 ByteArrayOutputStream 将急剧增长，不时重新分配其大小。相反，立即输出:

try {
    URL url = new URL("https://www.kaltura.com/p/....");
    InputStream in = new BufferedInputStream(url.openStream());
    res.setContentType("application/octet-stream");
    res.setHeader("Cache-Control", "no-cache");
    res.setHeader("Content-Disposition","attachment;filename=Smac_03_48.mp4 (iPad).mp4");
    res.setStatus(200);
    OutputStream out = res.getOutputStream();
    byte[] buf = new byte[8192];
    int n = 0;
    while (-1 != (n = in.read(buf)))
    {
        out.write(buf);
    }
    out.flush();
    in.close();     
} catch (Exception e) {
    e.printStackTrace();
}

当然，浏览器重定向会更快。