Tab navigation
The Integration Test concept
The concept behind integration test is pretty simple: if two software products are going to run together on the same operating system platform, they have to be tested together with a focus on their interactions.
When you're working with an operating system such as z/OS (which itself is comprised of numerous elements and features, and on which many other products must run and coexist) or Linux, the testing becomes considerably more complex—and considerably more important—but the concept is the same. All the elements and features of z/OS must work together and because nobody runs just an operating system, we have to make sure it works with a multitude of other products and functions.
At IBM, product testing is and always has been extensive. Integration test is not a replacement for other IBM test efforts. We still perform what we think of as three classical test phases: unit test, function test, and system test—you might know them by other names in your own environment. These test phases mainly focus on an individual product or function, or on a subset of products and functions. To further enhance quality, we built on the foundation provided by unit, function, and system test by adding an additional test effort—integration test. With integration test, there's greater emphasis on the customer experience, cross-product dependencies, and sensitivity to end users.
Integration Test - the final verification
System z Platform test is the final verification of a z/OS release prior to its becoming generally available to customers. If you'll allow us to borrow a famous phrase, the buck stops here. Our job as members of the System z Platform test team is to ensure that all the elements and features of z/OS work seamlessly together and can support true production, mission-critical work while also providing all the industrial-strength z/OS advantages you're depending on: reliability, availability, serviceability.
Our objective is to validate the System z software platform in a manner such that you, our customers, might use it. That means doing much more than just ensuring that products work together—it means understanding what it takes to introduce, implement, and effectively use new products and functions in a full production environment while thousands of end users are expecting constant availability. In other words, we want to experience for ourselves the challenge of integrating various solutions into a business environment. And we want to experience it first, so we can hopefully save you some pain!
We document the results of our testing (including any pain we've experienced and how you can avoid that pain) in our test reports. We provide tested examples, both in our test reports and here on our Web site, that span multiple products and have been run in what we call a pseudo-production environment.
Does all of this mean you will never again find a bug in an z/OS release? Well, nothing would make us happier! But that's a claim we cannot possibly make—we don't believe any software test organization could truthfully make such a claim about their product. The claim we can make is this: we endeavor to provide you with the most stable operating system platform available in the marketplace—z/OS.
Our test workloads
So, what do we mean by a pseudo-production environment? Well, we've set ourselves up to be as customer-like as possible. Our team is organized in a way that many IT shops are, and our team members take on the roles of system programmers, operators, database administrators, and so on. We focus on providing availability of applications to end users, and we pay attention to service level agreements and performance objectives. We look at the recovery aspects and behavior of our systems from an end user's perspective.
Our goal, like yours, is to have our workloads up 24 hours a day, 7 days a week (24 x 7). We have workloads that exercise the sysplex, networking, and application enablement characteristics of our configuration.
For details about our workloads, see the Our Workloads topic in our test reports.
Our test configuration details
For details about our test configuration, see the hardware configuration, software configuration, and networking configuration topics in our test reports.
About the System z Platform Evaluation Test Team
Hello! We, the members of the System z Platform Evaluation Test Team for z/OS and Linux Virtual Servers, welcome you to our Web site! Many of us have been members of this team since its inception in 1995, when we were known simply as Platform Evaluation Test, or PET.
From the outset, our primary focus has been to run a data sharing Parallel Sysplex on the System z platform in as much of a customer-like fashion as possible.
As the operating system platform has evolved over the years, from MVS/ESA to OS/390 to z/OS, our mission has also expanded and, as a result, we have developed somewhat of a dual personality. We continue our focus on Parallel Sysplex functionality and data sharing, and we also focus on many other elements and features of z/OS, some of which are not necessarily sysplex-oriented. This, coupled with our attempts to be both a customer and a test organization at the same time, provides us with just that extra bit of challenge that we need! But we figure that only helps us to understand how challenging it is for you, our customers, to run multiple hardware and software products and make them all work together—and that's an understatement.
Our team has grown tremendously over the years. We started out with only 12 team members and now we have 30 and growing! We also added Linux to the mix. As we began testing the entire suite of operating system elements and features along with our Parallel Sysplex responsibilities, we realized we had to divide into subteams, which are described in the following sections.
Base operating system team
The base operating system (OS) team is responsible for testing the z/OS system-related elements, features, and functions, which include ARM, DFSMS, EREP, HCD, JES2 and JES3, MVS (also known as the Base Control Program or BCP), RMF, SDSF, TSO/E, and z/OS UNIX System Services and file systems (HFS and zFS).
The base OS team is also responsible for:
- System automation, using IBM Tivoli NetView for z/OS and IBM Tivoli System Automation for z/OS
- Performance monitoring and tuning
- Storage management
- Using Health Checker to tune our systems with IBM recommendations
- Executing recovery scenarios, which include CPC failures, coupling facility failures, coupling facility structure failures, coupling facility link failures, and subsystem failures, all while continuing to run our workloads.
The base OS team also includes the testing of z/OS networking elements with other z/OS elements. The z/OS elements and applications, such as MQSeries, WebSphere Application Server, NFS, and z/OS Communications Server, are driven by workloads running on various platforms, including AIX, Sun, HP, Linux, z/OS, and various flavors of Microsoft Windows. The hardware from which the workloads run includes System z processors, Intel, Sun, and HP workstations, and POWER workstations, through switches and routers from Cisco and Extreme Networking. We currently run FastEthernet, Gigabit, and 10 Gigabit Ethernet networks, using the OSA-Express feature. This combination of of hardware and software provides us with a robust environment in which to exercise the z/OS platform.
Middleware team
The Middleware team supports the following products that are not elements or features of z/OS but that run on the z/OS platform: CICS, DB2, IMS, IRLM, MQ, RLS, and WebSphere Application Server. These products drive our mission-critical online transaction processing (OLTP) workloads, which are all Parallel Sysplex data sharing workloads.
The responsibilities of the Middleware team include:
- Migrating to new releases of each product
- Implementing new function, which is either dependent on a new z/OS function or which will enhance the integration test environment
- Working with the base operating system team to verify the recoverability of the products in the z/OS environment
- Regularly upgrading service with preventive maintenance
- Fixing problems and reporting problems to IBM Service for the affected product
- Providing product support for other integration test teams
- Filling the roles of support programmers and administrators
Security team
The Security team focuses on the security-related elements and features of z/OS, which include:
- Cryptographic Services
- Integrated Cryptographic Service Facility (ICSF)
- PKI Services
- System SSL
- Security Server
- Enterprise Identity Mapping (EIM)
- Network Authentication Service (Kerberos)
- RACF
- Tivoli Directory Server for z/OS (LDAP)
The team also focuses on other security-related products on the System z platform, such as:
- Encryption Key Manager (EKM)
- Encryption Facility
- Tivoli Federated Identity Manager (TFIM)
- Tivoli Access Manager (TAM)
For interoperability testing, the Security team also focuses on security-related products on non-System z platforms. These products include LDAP and Kerberos. The operating systems on which these products run include AIX, Linux, and Microsoft Windows, although not all of the products run on every operating system.
The responsibilities of the Security team include:
- System programming and product support for all of the security products
- Acting as security administrators for the zPET organization
- Building and maintaining expertise for the security products
- Exploiting the security products with other products, such as HTTP with System SSL
- Collaborating with the other test teams to exploit the security products, such as FTP with Kerberos authentication
- Maintaining awareness of the latest security products and solutions
Testware team
The Testware team is responsible for the strategy, design, and development of applications that drive our z/OS base and middleware software. This team consists of software architects, designers, developers, and application programmers working together to create, deploy, and maintain our mission-critical OLTP and exploit emerging technologies.
The mission of the Testware team is to:
- Focus on key customer, System z, and Integration Test business drivers using defined strategy methodologies.
- Develop a resilient business application infrastructure which provides a security-rich, agile, available, and recoverable environment that can handle planned and unplanned events.
- Continue to push higher volumes of transactions through our applications.
- Move our application suite from monolithic to composite applications using an SOA framework and development model.
- Standardize and streamline the application portfolio by eleminating redundancy.
- Reduce the cost of ownership to the business through efficient maintenance and operation of legacy applications.
Linux team
The primary responsibility of the Linux team is to emulate leading-edge customer environments, workloads, and activities. This includes cloning Linux images on z/VM and establishing security in a heterogeneous Linux server environment
In addition to the z/OS elements, the Linux team uses the following software in their day-to-day operations: z/VM, Linux on System z, WebSphere Application Server, WebSphere Application Sever Network Deployment (including Edge Components), Tivoli Access Manager for e-business (TAM), TAM WebSEAL, Tivoli Risk Manager, TrendMicro ScanMail, TrendMicro ServerProtect, iptables, z/OS LDAP and z/OS RACF, DB2, Apache plus System z hardware cryptographic acceleration, and various other open source security products.