The Making of a Data Center

A Data Center is a cost-efficient way to provide computing resources to a large corporation. Modern data centers are comprised of Mainframes, Unix and PC Servers, PC workstations, Tape Silos, Tape libraries,  High speed printers and bursters, DASD devices, and a 24-hour staff of people. As many as 100 people may be needed to maintain a large data center.

The role of the mainframe computer is central to current data centers. They provide the Computing power necessary to run the vast applications, manage large databases and move the vast amount of information required by a large corporation.
 
 

Automation Automation of Client Server, Workstation and Mainframe tasks which usually required a human to operate.  With software and hardware tools at our disposal we are able to Automate many tasks in the DataCenter.  Computer Operations  Computer Operations manages the UNIX application Servers and Mainframe Systems. Level 2 Support is provided to EHC. Data Management  The administration of Data at NBDC. HSM, SMS, SAMS, MVS Catalog Structure, DASD Maintenance, Y2K Backup and Recovery, and Tape Stacking.  Scheduling  They schedule and monitor approximately 95,000 jobs monthly on the Mainframe and Client Server processors. They also fix an average of 1,500 job ABENDS (ABnormal ENDS) monthly.

Conversion of all Direct Access Storage Device (DASD) to RAID-5 technology is complete. Both IBM RAMAC and Hitachi Data Systems 7700 DASD use RAID-5 data striping and disk sparing technology for constant data access. This results in no down time or data loss for single DASD failures.
 
 

The combination of new DASD and  TAPE devices require less than half the floor space they did just a few years ago.

A Data Center can use Mobius TapeSaver software to combine tapes that contain small data sets (less than thirty-three percent of tape used) on to a single tape. This will reduce the amount of tapes used and free up tapes for reuse by larger data sets. Save $200,000 per year. The TapeSaver saved this Data Center from buying approximately 66,000 tapes in 1997. The physical library, consisting of over 300,000 tapes, has relocated one floor up and now utilizes a new tape rack system. This new rack system allowed us to reduce the floor space requirements by 50%.
 

TAPE SILOS   Automatic mounting of tapes Hold thousands of tapes Minimize wait time for jobs that require a tape mount Facilitate archive of large amounts of data

Slot Addressable Management (SAM) software for use in our tape library. Instead of locating tapes in numerical order, SAM identifies and records the slot where a tape resides in the tape library. Using a hand-held radio frequency display, the operator can perform various daily functions. These functions include the pulling of expired and vault tapes. Returning a tape to the library requires the operator to scan the bar-coded tape and then the slot where they want to place it. The operator returns all tapes to one area of the library instead of placing returned tape across all areas of the library.
 
 
 

Hand Held Radio used to scan bar code on tapes.  The placement and removal for vaulting if necessary is automated.

SYSPLEX (combining multiple mainframes through data channels )

A parallel sysplex environment positions us for future technological change and gradual growth, devoid of any business unit dependencies on operating system architecture. This allows the business unit applications to view and utilize the hardware and software resources as a vast pool. The three field data centers have been involved with this new architecture since 1994 and have achieved the best online availability numbers ever. In addition, the parallel processing architecture has provided the ability to balance the workloads across multiple mainframe images. This allows for more aggregate MIPS (millions of instructions per second)  to become available. in case of increased activity.
 
 
 

These mainframe CPU's  run applications which update IMS , DB2 and PeopleSoft databases. These mainframes share loads based on which CPU has lighter load.

TOP

End-To-End Management.

Comprehensive control of the enterprise requires end-to-end management of the entire IT infrastructure. Enterprise management issues are not bounded by specific technologies. End-to-end management must address all types of resources: systems, networks, desktops, databases, and applications.
As IT grows to larger and larger systems, and as we expand the scope of end-to-end management, the view of the enterprise’s resources can become overwhelming.

Unicenter TNG integrates the management of all these resources and provides a cohesive view of this complex environment.

For example, a Unicenter TNG management function such as performance, deals with the performance of network devices and networks as a whole; systems and databases running on them; client/server applications and Internet applications. This integration allows it to provide a complete picture of the performance of all the IT resources involved in a business process. It lets administrators answer business-relevant questions: Are we able to process orders? Why is accounting slow? How secure is my customer information?

These Business Process Views enable administrators to deal with only those IT resources that impact each business process. By eliminating the data clutter that administrators face today, Unicenter TNG reduces the complexity of IT management.

Real World Interface.
Today’s enterprise environments are extremely complicated. End-to-end management involves enormous amounts of data. Gaining an intuitive understanding of all this information is a daunting task. Classical user interfaces are no longer adequate—we need a new approach. The revolutionary Real World Interface uses 3-D visualization and animation to display the entire IT environment. It provides an intuitive way of navigating among the systems and connections in the network, beginning in the cities and buildings of the real world. Unispace, a representation of what is going on in the systems, visualizes abstract objects such as processes, databases, jobs, and users.  The Real World Interface allows for a style of interaction that is intuitive and representative of the way people naturally solve problems.

CA-OPS/MVS II is a comprehensive automated systems operations product that maximizes system availability, offers improved efficiency, reduces errors and downtime, and increases productivity at all levels. CA-OPS/MVS II is the leading automation product in the mainframe market due to its dependable and efficient architecture, a powerful automation language, and various facilities geared at enabling rapid automation development. CA-OPS/MVS II plays a key role in Computer Associates' overall enterprise automation solution as a powerful standalone tool for ensuring maximum availability of the MVS platform, as well as extending manageability and visibility in the enterprise.

CA-View is an automated archival and retrieval system that allows immediate online viewing of computer output. It is tightly integrated with CA-Deliver to provide a total MVS output management solution. With CA-View end users can view reports in a user-friendly environment, printing only those pages they want. The logical viewing facility allows users to customize their reports, eliminating unnecessary columns and lines. The customized view can also be printed. With fewer reports printed, less time is spent reformatting, tracking, and rerunning reports. In addition, CA-View supports direct read capabilities from optical, tape, and robotics devices. These interfaces, along with extensive indexing capabilities, provide customers with a comprehensive Computer Output to Microfiche-Replacement (COM-R) solution. CA-View works with our PC client component, CA-View Workstation, extending report viewing from the mainframe to the desktop, using a Windows Graphical User Interface (GUI). By adding CA-Connect to this environment, unattended downloading of reports to the workstation user is available from CA-View. Additionally, adding CA-DocServer provides a total solution to document management in the open systems environment. This addition allows the automated flow of reports between MVS and LAN platforms, offering a solution set that grows with changing business needs.

Unicenter TNG Automation Point Option features extensive automation, problem notification and escalation capabilities and is integrated with Unicenter TNG. As an integrated part of Unicenter TNG or as a standalone product CA-Automation Point provides an outboard automation solution for IBM and non-IBM systems. Events can be consolidated from multiple platforms for correlation and cross-platform management. If a system problem occurs, CA-Automation Point can automatically correct it and can even respond to irregularities within the data center infrastructure (e.g., environmental equipment such as air conditioners, generators, etc.).

TOP

Compiled by Mark Wittman, from a local corporation's Data Center newsletters.