Chapter 5. Case study
5.4 Identifying the reports requirements
One of the most important steps to be performed when planning to either migrate to or deploy Tivoli Decision Support is to gather all the customer reporting requirements. This will enable us to decide, for example, which TDS Discovery Guides will be used, how many Tivoli Discovery Administrator machines will be needed, and what the required hardware configuration for each component of TDS will be.
In this section, we will discuss and identify both the customer’s requirements for reporting and all reports that are currently being used to satisfy these
NCO
requirements. Future requirements, as well as future recommendations, will be discussed in Section 5.10, “Future reporting requirements” on page 162.
5.4.1 Customer reporting requirements
Reporting requirements in Distributed Systems Management (DSM) translate into five fundamental metrics:
• Availability Reports
• Performance Measurement
• Response Time to Failures (SLA)
• Cost Prediction and Measurement
• Detailed Applications Related Reports
5.4.2 The SDC actual solution for reporting
As shown in Figure 39 on page 94, IBM SDC West uses a variety of platforms and systems-management tools to collect the required data to generate reports of its distributed environment. The challenge is to find a solution that is able to process and interpret the data stored in different databases in different servers.
It is beyond the scope of this redbook to identify all the reports generated and used by IBM SDC West. We will only consider the reports that cover the basic metrics of availability, capacity and performance, response time to failure (sla), and cost if available.
Note
Figure 39. The Problem for reporting
IBM SDC West uses two methods for reporting. Despite valiant efforts, these two methods are still facing the problem of using multiple tools, dealing with different sources of information, and storing the data in files with different formats. The first one will be referred as the in-house method, which has been developed by the IBM Service Delivery Center Tucson, Arizona. The second method, Server Resource Management (SRM), has been developed by the IBM Global Service South Performance & Capacity team.
For organizational and security reasons, both methods utilize the concept of accounts to access the reports. Each group of people responsible for their account has the capability of looking only at the reports that are related to their business and interest. There are IBM internal accounts, which are related to IBM internal departments or locations, and external accounts, which are related to IBM customers. All these reports can be reached either through the IBM Intranet or Internet. For the purpose of this case study, we will look at the reports of an internal IBM account called Network Computing Offerings (NCO). The NCO account uses both methods of reports. We will identify the reports available in the actual SDC solution for reporting and then map these reports with those produced by Tivoli Decision Support.
5.4.2.1 The in-house method
The in-house solution relies on many sources for collecting data, such as:
• UNIX-based tools:
Analysis
Reports
. . .
Tool 1
Tool n Tool 2
• Thevmstatcommand is sampled every minute and used to compute an hourly average.
• Process memory is checked witht theps gvcommand and paging space percentage full is checked with thelsps -a command and sampled hourly.
• Network packets in/out and errors in/out are sampled daily using the netstatcommand.
• File systems snapshot is checked with thedf -kcommand and sampled daily.
• Tivoli Applications:
• Tivoli Distributed Monitoring
• Tivoli NetView
• Tivoli Enterprise Console
• Netfinity Capacity Manager
Figure 40 shows how the process is used for reporting the in-house method:
Figure 40. The in-house process for reporting Tivoli
A pplications
N e tfinity M on itors A IX Too ls
Phase 1 - D ata Co llection
Phase 2 - D ata Processin g
Phase 3 - R eport G eneration M onitors
Pe rl Scripts and Program s
HT M L and Ja va form at
IBM Intra net
The three phases of the in-house process are detailed in the following list:
• Phase 1 - Data Collection
The monitors collect relevant information according to some predefined thresholds and write the data to files on a file server.
• Phase 2 - Data Processing
As soon as the data is collected, it is processed daily by the in-house Perl scripts and programs populating some flat files for each server and metric.
The rolling year of data is kept.
• Phase 3 - Report Generation
The reports are generated in HTML and Java format showing each month's data in tabular format with links to applets that graph the data sets for the year by account (server group).
5.4.2.2 The SRM method
In order to satisfy the actual requirement for reporting, IBM Global Service South Performance & Capacity team has developed a set of Server Resource Management tools to expand the performance and capacity trending on the Distributed Systems Management (DSM) platforms and applications, such as AIX/UNIX, Windows NT, OS/2, SAP R/3, and Lotus Notes. The SRM tool set is used for both internal and external account reporting.
The SRM solution relies on existing monitoring tools for each available platform, such as Netfinity Capacity Manager (CISC platforms), Perl Scripts (RISC platforms), Zperstat (for SAP R/3), and NotesView (Lotus environment) to collect and pass on event data.
The SRM tool is divided into four main components that enable the
performance and capacity trending process as detailed in the following list:
• Phase 1 - Collection
The data is passively collected in multiple formats and stored into files.
• Phase 2 - Transmission
The SRM tool receives the data and converts it to a common format. Then the data are stored in a single database in a convenient format.
• Phase 3 - Analysis
An automated process provides Distributed Systems Management resource trending and exception analysis in this phase.
• Phase 4 - Web Preparation and Presentation
The data is processed and HTML and Java reports are generated and published on the IBM Intranet.
Figure 41 graphically represents the SRM collection and reporting process:
Figure 41. The SRM method for reporting
5.4.3 The Reports of the NCO account
In this section, we will show the reports generated by the two methods described in Section 5.4.2, “The SDC actual solution for reporting” on page 93. The goal here is to provide information about the current scenario of reporting used at IBM SDC West to satisfy the actual requirements.
5.4.3.1 In-house reports
The following are the reports available for the NCO account using the in-house method for reporting: