IBM的RDZ如何在PC上运行，访问大型机的文件系统？ [英] How does IBM's RDZ running on a PC access the mainframe's file system?

问题描述

我一再听到，尽管IBM大型机上可以使用NFS风格的文件系统，但是它们往往没有启用，可能是为了尽量减少大型机相对于世界其他地方的安全风险。

b
$ b

考虑到我想生产基于PC的工具，在大型机上处理和处理文件，这会产生一个简单的问题（打开NFS文件\ mainframe \ foo' ）更难;我能指望在网络环境中提供文件系统访问吗？
（Linux系统通过Samba提供NFS几乎是标准的，所以这很容易）。
$ b IBM提供了Rational Developer for Z，一个由IBM使用的Eclipse变体COBOL程序员。
RDZ似乎可以直接访问IBM大型机文件系统。他们用
来做什么？为什么不能提供给我，如果是这样的话，是什么？

守护进程（在UNIX中说），它在z / OS主机上运行并接受来自Eclipse插件的连接。该协议是专有的，所以你不可能找到任何有关它的信息。

而RDz不仅仅是COBOL程序员。它被用在许多商店里，人们希望在大型机上存储它们的源代码 - 为什么要维护两个独立的存储库？这就是为什么它有那些长名称/短名称和ASCII / EBCDIC翻译来将那些不敬虔的Java路径变成我们精美优雅的8字符成员名称，并允许我们在z / OS下读取它，虽然ISPF编辑器的源ascii命令缓解了最后的担忧。

如果你想做类似的事情，你需要编码自己开始的任务来接受来自客户的传入连接。这听起来不那么难。因为USS（UNIX系统服务，更名为OpenMVS）作为基本操作系统软件的一部分附带了z / OS，所以您将在UNIX环境中实际执行此操作。它允许你透明地访问USS文件和z / OS数据集/成员。

然后，你需要说服大型机商店，你开始的任务是不是安全风险。让我知道这是怎么实现的： - ）

您可能会发现将NFS作为软件的先决条件会更容易。那么，至少，这是IBM的安全问题，不是你的。

I hear repeatedly that while NFS-style files systems are available on IBM mainframes, they are often not enabled, presumably to minimize the security risks of the mainframe vis-a-vis the rest of the world.

Given that I'd like to produce PC-based tools that reach out and process files on the mainframe, this makes a simple problem ("open NFS file '\mainframe\foo'") much harder; what can I count on to provide file system access in a networked enviroment? (Linux systems offer NFS via Samba pretty much as standard, so this is easy).

IBM offers Rational Developer for Z, an Eclipse variant used by IBM COBOL programmers. RDZ seems to have direct access to the IBM mainframe file system. What are they using to do that? Why isn't that available to me, and if it is, what is it?

解决方案

RDz has a started task (a daemon in UNIX-speak) which runs on the z/OS host and accepts connections from the Eclipse plug-in. The protocol is proprietary so you're unlikely to be able to find out any information about it.

And RDz isn't just for COBOL programmers. It's used in many shops where people want to store all their source code on the mainframe - why maintain two separate repositories? That's why it has those longname/shortname and ASCII/EBCDIC translations to turn those ungodly Java paths into our beautifully elegant 8-character member names and allow us to read it under z/OS, although the ISPF editor's "source ascii" command has alleviated that last concern somewhat.

If you want to do a similar thing, you'll need to code up your own started task to accept incoming connections from your clients. This isn't as hard as it sounds. You'll actually be doing it in a UNIX environment since USS (UNIX System Services, the renamed OpenMVS) comes with z/OS as part of the Base Operating System software. And it allows you to access both USS files and z/OS datasets/members transparently.

Then, you'll need to convince the mainframe shops that your started task is not a security risk. Let me know how that works out for you :-)

You may find it easier to just make NFS a pre-requisite of your software. Then, at least, it's IBM's security problem, not yours.

这篇关于IBM的RDZ如何在PC上运行，访问大型机的文件系统？的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！