项目经验-李强

项目一：SEMEA

项目简介（功能与用途）：

IBM Global Service, the largest department of IBM world wide which contributes nearly half of IBM’s annual income every year, has made a long-team strategic decision. China, one of the most rapidly developing countries is the chosen one of the outsourcing centers for IBM Global Service business. IBM EMEA(Europe, Middle-East & Africa), one of the three continent made up IBM, is delivering some of its strategic outsourcing business to ISSC Shenzhen, including application development, system administration, project management, security & auditing…

SEMEA mission is one of the submission for the EMEA project, mainly focusing on south-European clients including: Banks, Telecoms, Large Manufacturers, Governments. The hardware are located in different countries. We handle the various data center businesses via remote control. Roughly summed, The mission is made up of 7 mainframe severs, 15 LPARs, 25 DB2 subsystems for porduction, 1 sysplex for application/test system. More details including hundreds of databases, thousands of tablespaces, over two hundred partitioned table spaces, nearly one hundred LOBs, hundreds of tables with over 1,000,000 records, over three dozen of Tigabyte tablespaces.

 

项目难点与解决方案：

1.       DB2 Health Check:

   When we think about software health, we think about efficiency, code path length, speed, lack of bugs, and so on. In a broader sense, we also consider how easy new features are to use, whether new functions perform well, and how generally fault tolerant is the system.

Overall health, therefore, is a combination of the following:

Efficiency — raw performance

Flexibility — ability to handle and exploit features and functions

Fault Tolerance — ability to tolerate failures

Further, there are several components in a DB2 subsystem. They include: Software, Process, People

So, an overall health measure for a DB2 subsystem consists of a combination of these three characteristics of health across three categories of items. For more details please refer to “ATTACHMENT1”

All the “ATTACHMENT” are listed from page 11

2.       Key points in DB2 Version 8 transition:

   Transition to DB2 V8 is not as easy as submitting JCLs! V8 has made much greater improvement than old version and you’d better know each difference before get into start. As more shops embark on the transition from DB2 Version 7 to V8, it’s time to review some of the biggest things to “keep an eye on” as you migrate your DB2 subsystems. DB2 V8 includes a substantial number of changes. There is a book called “DB2 UDB for z/OS Version 8 Everything You Ever Wanted to Know” which can give you a complete view. But it needs your patience to read over one thousand pages. Here summarizing from uncountable happy and sad real stories, I’d like to share 5 biggest transition considerations which are worth keeping in mind at first: please refer to “ATTACHMENT2”

 

项目成功与失败的经验归纳：

1.       Success: DB2 subsystem migration on mainframe:

According to customer’s demands, we need install new DB2 subsystem or migrate to version 8 with minimum effection on current data and system performance during limited hours. Carefully planning and fully investigation is key for success. To write documentation with your own experience can also help you replicate the process to other LPARs so as to minimize the whole down-time. Here I’d like to show some of

 

 

my working diarys as solutions:

   Sample Migration checklist to DB2 Version 8 please refer to “ATTACHMENT3”

 

2.       Failure: How to find a efficient way to manage DSNZPARM

   DSNZPARM, DB2 run-time configuration lib including all the basical system parameters, is one of the key object to minor and manage for DB2 system programmers. It is originated from the installation panel and link edited by the installation job: DSNTIJUZ. Unfortunately the job/panel is not properly secured at the beginning, although you could use system level security tools such as RACF to protect the datasets from been accessed by the users outside the DB2 group. It’s still hard to safeguard them from being abused by DB2 users themselves. Take an example, you made some changes to ZPARM according to customer demands, since you are the install-sysadmin you do not think it necessary to tell other team members. But one day, you leave for vacation and your backup comes, if you don’t tell him what you’ve done to the whole system before, he could know nothing about the change but consider the ZPARM as the default one. Worse more, he may also update the ZPARM during your absence. When you come back and he forgot to tell you, thing would entirely out of control

   Learned from the tears, currently we have a well-round plan for DSNZPARM, we made a special library for all the historical ZPARMs, Before updating it to new version, you must rename the in-use one to ZPAR××××(×stands for the date we drop it), then submit job to generate a new one with the exact name DSNZPARM. Further more, we also create a new library for historical job DSNTIJUZ. Inside the JCL, you must write the date, name, and reason as JCL comments near the changed line. If JUZ1 is the original one and you want to update ZPARM to new version, you must use JUZ1 as template to create JUZ2 and submit.

Therefore both the DSNZPARM lib and the DSNTIJUZ lib will keep a historical track to control their versions.

 

你在项目中岗位与贡献：

Database system programmer

Duties include:

Responsible for day-to-day DB2 systems programming support to IBM client or any other client during Sydney business hours.

Perform DB2 OS/390 installation/migration/maintenance/support/customisation of Db2 parameters

Install/configure QMF, DB2 administration Tool, DB2PM, Omegamon for DB2 and other similar products and DB2 Utilities

Provide rostered on call support.

Perform assigned simple to medium DB2 problem diagnosis/resolution  tasks

Implement assigned changes following specific documentation

Perform security compliance related tasks against specific documentation

Identify and recommend improvements to existing processes and procedures

Perform assigned simple/medium Problem and change management tasks using client problem/change management application that will be assigned and delegated by the IGS EMEA team leader.

Other activities as directed by the IGS EMEA manager or team leader

 

 

项目二：Deutsche Bank

项目简介（功能与用途）：

Deutsche Bank is a leading global investment bank with a strong and profitable private clients franchise. € 1.035 billion in assets; 64,103 employees from 130 nations; Unparalleled financial services in 73 countries--A leader in Germany and Europe, the bank is continuously growing in North America, Asia, and key emerging markets.

IBM Global Service has taken over its IT operation and DB is now one of the biggest client for IBM Germany and also EMEA, some of the datacenter business is now deliver to China as another mission of the large EMEA project. We are focusing on DB2 system administration.

Compared with SEMEA, most of which are various departments inside IBM, DB is a stand-alone client at all, therefore, it requires much more high-level Service Level Agreement, including:

1.Higher service hours, 24*7 with less down-time allowed

2.Higher security level with more strict auditing process

3.Lower response time required. Customer’s requests should be replied by no means of more than 20 minutes.

 

项目难点与解决方法：

1.       Hardware and Software Solutions for Mainframe Database Recovery:

Our customers have invested millions of dollars in their mainframe database applications. These applications allow trains and planes to move, packages to be shipped, financial transactions to occur, and manufacturing to proceed. If these applications are unavailable for any reason, the company is not receiving the expected gain on their investment. Further, some industries receive fines if certain transactions are not processed in a timely fashion. Recent legislation requires all publicly traded institutions to maintain a high-availability application plan, including recovery from total site disaster.

For more details, please refer to “ATTACHMENT 4”

 

2.       DB2 Buffering:

   Memory usage is one of the most important areas for tuning DB2 subsystem performance. DB2 for z/OS uses memory for buffer pools, the EDM pool, RID pool, and sort pools to cache data and structures in memory. The better memory is allocated to these structures, the better DB2 will perform.

   During the years working as an DBA for DB, I have made some notes about BP tuning. Since internal articles about DB2 BP tuning seems not as much as those for oracle, I’d like to share my experiences in this field, please refer to “ATTACHMENT5”

 

项目成功与失败的经验归纳：

1.       Planning for disaster recovery:

   In a true disaster, how can you get DB2 up within minutes and with minimum data loses, As your enterprise, like DB, response to the availability demands of e-business, the current standard of 24*7 service level are not ask for your carefulness and wisdom to build a Disaster Recovery Planning (DRP), according to the survey from IDUG, less than 50 percent of companies responding to have a DR capability. Of those that do, usually fewer than 50 percent have ever tested their plan, which is tantamount to not having one at all. Here I’d like to show part of our DRP as a reference. Please have a look to ATTACHMENT6

  

2.       The lessons we leant when disaster really strikes:

   Preparing for disasters is one thing; experiencing them is quite another. Under severe stress and pressure, businesses do what they can to stay online and ensure employee safety. Nevertheless, no matter how well you plan, something’s always overlooked. Disasters will always present surprises, but preparation favorably positions sites to respond to them. Following is a list of “lessons learned” from sites participating in a sudden system failure at DB Frankfurt last year, and from disaster recovery experts in IBM Germany:

• Don’t wait to respond to disaster “threats.” Key personnel for disaster recovery and business continuation should be dispatched to hotsites and other alternate locations before a storm or other forecasted risk makes it impossible or difficult to leave.
• Understand the Recovery Time Objective and Recovery Point Objective (RTO/RPO) for your business, and make sure you have system processes, facilities, staff and tools in place to meet them. Regularly test your DR plan.
• Strongly consider staging a fully integrated, total test of your DR plan instead of isolated pieces. This lets you identify areas of exposure you might have missed.
• Make sure your DR documentation is current. Old documentation can be worse than no documentation.
• Automate critical data set identification and handling and the overall recovery procedure as much as possible.
• Regularly participate in tabletop exercises with your key executives and IT staff. Everyone has their DR plans and updates, but few people have walked through them.
• Train your employees in disaster preparedness—not only for work, but for home.
• Remember: People are the most critical issue.

 

你在项目中岗位与贡献：

DB2 DBA

Provide day-to-day Database Administration support during German business hours.

Perform assigned  Problem and Change management tasks using standard tools and ad hoc tasks/service requests that will be assigned and delegated by the IGS EMEA team leader

Perform assigned simple to medium DBA problem diagnosis/resolution  tasks

Support Business Performance Metrics (BPM) requirements

Provide DBA support and testing on DB2 and DB2 related product upgrades

Implement assigned changes following specific documentation

Provide database security management in DB2 as requested

Perform security compliance related tasks against specific documentation

Perform normal DB2 DBA functions such as backup, recovery, reorganisation, runstats etc Provide DB2 database support in non production environments such as create/alter objects, data refresh etc. Ensure support documentation is updated when new database objects are created as appropriate per team process

Perform health check of the databases according to the daily checklist

Monitor closely SMS storage pools used by IMS/DB2/CICS database datasets in all mainframe environments, perform DASD forecast quarterly and raise requests for additional DASD to System Storage Group

Perform other activities as directed by the IGS EMEA manager or team leader

 

 

 

 

项目三：Telstra

项目简介（功能与用途）：

Telstra is Australia's leading telecommunications and information services company, with one of the best known brands in the country. They offer a full range of services and compete in all telecommunications markets throughout Australia, providing more than 10.3 million Australian fixed line and more than 6.5 million mobile services.

Telstra has outsourced its IT business to IGS Australia, and become the largest client for IBM Australia and Asia Pacific, another one of the three continents of IBM global. Spreading all over Asia Pacific, IGS A/NZ has a huge cross-nation development and administration team, including the applications developers, software architects, business analyzers, system administrators, security auditors, project managers, global managements, end-users, and so on. I was chosen as the performance analyzer for its cross-national development team this year, in charge of the application performance analyzing and tuning

 

项目难点与解决方法：

1.       Terabyte Data Warehouse Table Design:

   Data warehouse tables continue to fill up at amazing speeds with larger and larger amounts of data. This situation has persisted since the beginning of technology. Now, with all the advances in DASD and with DB2 Version 8 there are many design choices. The following will discuss some of the options for designing tables for terabytes of data.

Regardless of whether the data will arrive at one time or over a long period of time, there are many concerns that need to be addressed; but, usually, query access and maintenance are the most important. Query access means being able to get at the information you want immediately through an indexed access or by scanning only a small portion of the data. Maintenance means reasonable backup and/or recovery costs and never having to reorganize the data because of its size or the maintenance downtime issues.

For more info, please refer to “ATTACHMENT7”

 

2.  DB2 Subsystem Tuning in an ERP Environment

  The most important issues facing the systems programmer when configuring a DB2 subsystem to successfully support an ERP package installation are:

Catalog / Directory Management

Log Management

DBM1 Address Space Management

Dynamic SQL Caching

Virtual Pool Selection and Sizing

  Details for each item are described in “ATTACHMENT 8”

 

项目成功与失败的经验归纳：

1.Irregular Sorting Requirement:

   Sometimes the requirements of a particular application dictate that data needs to be sorted using some irregular collating sequence. These odd needs sometimes cause developers to sit and scratch their heads for hours, searching for ways to make DB2 do something that seems “unnatural.” But often you can create an answer just by understanding the problem and applying some creative SQL.

At this point, some of you might be asking “What the heck is he talking about?” Fair enough. Let’s look at my success stories to clarify the issue. Refer to “ATTACHMENT9”

 

2.       Change Management Standards:

   For large projects with various job roles, technical skills may be not the key point to success. What really important is the discipline identifying each role and each steps for all the processes. Especially for cross-nation developing groups, team members living in different country are hard to face each other at all. It’s the disciplines that make us together as a whole team and clear our targets and methods to avoid conflicts/puzzles. Not only management guys but also technical guys should have a strong sense for it. Unfortunately I didn’t take enough attraction on process managements at the beginning. I used to waste a lot of time negotiating to clients, not service managers. As a result, I always felt depressed and confused for customers seem do not listen to me at all. Thanks to my manager who helped me out of the puzzle, since then, I’ve been working not hardly but smartly. Here comes some of the Change management standards I’ve learned for your reference:

Close   

Change  

Execute   

 Change    

Approve  

Change  

Assess     

Risk 1 or 2  

Change     

Schedule   

Change    

Check Request

 and create  

Change Plan  

Request

and Assign

Change

Requester

Monitor and Control

Job Roles and Responsibilities

Assignee

Coordinator

Controller

Assessor

Approver

Requester

Installer

Activator

All

Competencies

All Delivery

Competencies

Service Management

All Delivery

Competencies and DPE/SDM

All

Competencies

All Delivery

Competencies

All

Competencies

Coordinator

Controller

Service Management

Process

Role

Executed

by

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Close   

Change  

Execute   

 Change    

Approve  

Change  

Assess     

Risk 1 or 2  

Change      

Schedule   

Change    

Check Request

 and create  

Change Plan  

Request

and Assign

Change

Monitor and Control

The Critical Process Steps

Check Request

•Risk Category ok?

•Approver list ok?

 

Change Plan

•Test

•Implementation

•Backout

•Impacts

 

 

Formal Check

•Risk Category

•Approver list

•Change Plan

•Timeframes

Assessments

•Business

•Technical

Approval

•In time

•List complete

•DPE = last approver

 

Implementation

•fully approved

•Keep timeframes

•Consider Backout

Implementation

•Verification

•Customer info

•Correct flag

    e.g.  ok / nok

• Change Categorization

• Deciding about assessment and approval

• Obtaining approval or endorsement

• Run a Change Advisory Board

• Identify and execute necessary escalations

• Analyze process and identify places for improvements

你在项目中岗位与贡献：

 

Application Performance Analyst

Duties include:

The primary objective of the position is to provide Application Performance Analysis and Reporting as required for various applications to meet client system specifications.

* DB2 SQL tuning

* Application Performance tuning

* Application design review

* Performance Impact Assessment (PIA) and reporting

* Online Performance monitoring

* Application performance problem investigation and resolution

* Consequential Impact Assessment (CIA) – provide input and review

Work closely with AMS teams to build and implement application releases 6 times a year.  Staff will attend Australian led team meetings during Australian business hours.

Perform assigned Problem and Change tasks using Telstra’s standard tools and ad-hoc tasks/service requests as assigned and delegated by PCM Manager of team leader. The problem and change queue in Solve must be checked and actioned regularly

Work closely with AMS teams to build and implement application releases 6 times a year.  Staff will attend Australian led team meetings during Australian business hours.

 

 

 

 

 

项目四：SEG & SAMSUNG

项目简介（功能与用途）：

SEG & SAMSUNG, the joint venture for SEG Shenzhen and SAMSUNG KOREA is one of the biggest exporter in Shenzhen. Comparing to the rapidly growing business, their old ERP system developed in 1980s is going to be out of date. Conducted by my mentor from IBM case laboratory, we are ask to update parts of the database structures and re-code some programs. It was my first job when I was a student, which drive me into the database field.

 

 

项目难点与解决方法：

Understanding Customer demands

Server years ago, as a famous programmer inside campuse, I’ve spent months improving my technical skills so as to be confident enough to take over the mission. However, It turns out a severe defeat at the beginning. I can not understand what are the managers and directors talking about, needless to say the logical and physical design!

The reason for misunderstanding is we have none knowledge on customer’s business at all in despite of technical skills we acquired. With such understanding, we spent days and nights studying customer business. We learned from all roles in the firm, operators, security guys, technicians, engineers, clerks, managers…, read all the documentations and papers. After nearly half a year investigation, we received a happy result when knocking at the presedent’s door again.

 

项目成功与失败的经验归纳：

Understanding customer demands:

1.       Basic knowledge for customer business, the more the better

2.       Dedication to customer: Sometimes as a technical expert, you may feel rather depressed for customer’s rejection to your wonderful design, even worse, customer may insist on their “stupid” solutions. However, don’t stick on your point. Why not considering from your customer’s view? Perhaps, you two are just concentrating on different points. For a certain reason, customers may not like to tell you their deeper concerns. Although their solution is stupid on technical, they may be the best plan considering the hidden situations you don’t know.

 

你在项目中岗位与贡献：

Lead our team to customize client demands and re-code some ERP programs before dead line.