|
This is part ten of the series that began with
What is System-Managed Storage?
The Challenge
Information processing is used not just to keep up with the competition, but to outrun it. The question is, for a given problem, who can provide the best solution: IBM, your company, another company, or a combination? Many ready-made solutions exist, many of them within the UNIX world. Once you invest in one, you need it to keep running until it can be replaced gracefully.
In response to these needs, IBM is using three dynamic building blocks to create an information environment for the future.
The first is today's powerful workstations. As workstation capabilities increase, users become increasingly self-sufficient. Yet they must work with other users as part of a network. Linked today with LAN servers and "groupware," tomorrow's users will need access to even more resources. They will expect them to be managed elsewhere, freeing users to be more creative and productive.
The second building block is open systems, often designed around UNIX standards. Open systems promise the freedom to design an environment using a combination resources suited to your needs with the assurance that your investment in applications and data will be portable to the new configurations of the future.
For many businesses relying on the strength of one operating system for the critical information processing one more building block is needed: MVS. MVS offers large-scale, high-performance transaction, interactive and batch computing on a scale unmatched by smaller, less sophisticated operating systems. Among its strengths are:
- The ability to handle many interactive users
- Efficient sharing of resources
- Sophisticated backup and archival facilities
- Thousands of application solutions
- Extensive security and auditing features
- Recoverability and system integrity features
- An outstanding record of reliability and service
What is OS/390 UNIX System Services?
OS/390 UNIX System Services (OS/390 UNIX), formerly known as OpenEdition MVS, is a part of IBM's open and distributed strategy for MVS. It combines the power of the workstation, the flexibility of open systems, and the strength of MVS. Offering open interfaces for applications and users, OS/390 UNIX supports and fosters an environment of larger operating systems or servers and of distributed systems and workstations that share common interfaces. With OS/390 UNIX, you have the industrial strength of MVS while gaining access to applications and data in multivendor systems.
The open and distributed strategy views MVS as both server to and peer with other operating systems in a network. OS/390 UNIX extends support to the set of standards being developed for POSIX.
Applications that conform to the POSIX standard for C programs can be moved with relative ease among conforming systems and can meet national standards and guidelines. With OS/390 UNIX, your investment in MVS applications and data is protected and enhanced, and new applications have access to existing data.
Using OS/390 UNIX, users can switch back and forth between traditional TSO/E interfaces and the OS/390 UNIX shell interface. This offers familiar commands and interactive menus to create and manage hierarchical file system files and copy data between MVS data sets and files to both MVS-skilled users and UNIX-skilled users. Application programmers and users can use both sets of interfaces and even mix them.
The Hierarchical File System
OS/390 UNIX creates a hierarchical file system within a new kind of MVS data set: a hierarchical file system (HFS) data set. HFS data sets, provided by DFSMS/MVS, contain the HFS file structure. This structure is a framework of directories and files called a file system. HFS files are identified and accessed using the POSIX-compliant OS/390 UNIX facilities.
An HFS data set is allocated by specifying a DSNTYPE of HFS either explicitly or using a data class. HFS data sets must be system-managed, and an individual HFS data set can reside on only one DASD volume. Like PDSEs, HFS data sets can expand to up to 123 extents of DASD space.
With DFSMS/MVS, OS/390 UNIX manages HFS data sets, providing access to the files they contain, as well as backup/recovery and migration/recall capability. HFS data sets:
- must be system-managed, reside on DASD volumes, and be cataloged
- must be accessed using POSIX system calls
- are supported by DADSM create, rename, and scratch
- are supported by DFSMShsm if DFSMSdss is used as the data mover
- are not supported by IEBCOPY or DFSMSdss COPY
OS/390 UNIX provides facilities to make accessible or inaccessible subsets of files in any hierarchical structure by logically "mounting" any particular hierarchy. A particular hierarchy that has been mounted can later be "unmounted."
DFSMS/MVS works with OS/390 UNIX to provide a full UNIX environment within MVS/ESA. It is compliant with applicable Universal UNIX and X/OPEN standards, enabling you to have an open system with the capacity, reliability, and storage management capabilities of MVS/ESA.
Besides protecting the hierarchical file system data set itself, OS/390 UNIX also can protect the individual files within the data set. The capabilities of MVS/ESA and related products such as RACF allow you to:
- Limit and control file access to specific users
- Define procedures for automatically backing up and migrating file systems
- Define a policy for the storage placement and use of file systems
Backing Up File Systems
DFSMShsm can automatically back up HFS data sets using DFSMSdss as the data mover, but it cannot back up individual files within an HFS data set. To back up individual files, you must use the ADSTAR Distributed Storage Manager (ADSM) client for OS/390 UNIX to perform back up to an ADSM server or manually issue OS/390 UNIX shell commands.
You can back up mountable file systems by periodically backing up the HFS data sets that contain them. DFSMShsm allows you to migrate and restore unmounted file systems. You can also manually back up a mountable file system with DFSMSdss data set dumps using the DFSMSdss DUMP command.
Migrating File Systems
If a file system remains unmounted for a predetermined time, DFSMS/MVS can migrate it to a lower priority storage medium. DFSMS/MVS automatically recalls a migrated file system from migration storage if a mount command is issued for the file system.
Accessing File Systems
DFSMS/MVS Network File System enables authorized NFS clients to remotely read, write, create or delete OS/390 UNIX files. When NFS clients access OS/390 UNIX files, the files on the MVS system appear as local directories and files to the client system. When coupled with the DFSMS/MVS Network File System, MVS/ESA becomes a full-feature UNIX server.
Data Exchange
With OS/390 UNIX, MVS becomes a truly open system with the ability to exchange data between the hierarchical file system and other systems. Data from outside MVS can be converted into EBCDIC and stored in HFS files where it can be accessed by both OS/390 UNIX applications and distributed applications. Data from conventional MVS data sets can be copied into HFS files where it becomes available to portable UNIX applications. In addition, applications written for MVS can access both OS/390 files and conventional MVS data sets.
In Summary
OS/390 UNIX and DFSMS/MVS can:
- Make your enterprise more competitive by allowing you to integrate applications from different sources and put applications and data in the best place in your network
- Help your enterprise to grow by capitalizing on the strengths of MVS while allowing you to take advantage of open standards
- Make end users and application programmers more productive by using features available through OS/390 UNIX
- Protect your investment while providing a path for future growth, by allowing you to run existing MVS applications unchanged and use POSIX for greater flexibility
[
Previous Article |
Next Article |
All Articles ]
| 
