CLEVER Solutions Empowering Global Enterprise Linux on System z Migration Case Studies Welcome to the latest edition of the AES Newsflash! Cloud is the top project for 2013 and will continue massive installations in 2014. Many of these involve Linux on System z. There are many aspects to a successful migration including organizational, application considerations, planning, performance and availability, diagnostic, disaster recovery and operational functions to name a few. In this issue, Laura's Corner covers pro-active problem resolution that you don't want to miss! Explore CLEVER solutions and services for improving Linux on System z migrations and reducing costs. Sounds like a CLEVER move you can't afford to miss! AES on the Road AES will again be at SHARE in Anaheim occurring March 9-14, 2014. Details will follow on specific AES activities. AES speakers David Cheng and Laura Knapp presented 7 sessions at the recent SHARE meeting in Boston. Sessions covered a wide range of topics from IPv6, IP Mobility, advanced IP diagnostics, and IP Management solutions. If case you 1
missed these sessions you can download the session proceedings from the SHARE website. Lauraʹs Corner Avoiding Problems on Linux on System z It is always refreshing when I receive questions and comments between distributions of Laura's Corner. In fact this months' discussion is based on questions I have received from many of you reading this email newsletter. One of my first managers taught me never to 'assume' and it is clear from the case presented that we can all use a refresher on performance management and availability management 101. Case Study This network diagram was sent to me by one of our clients and it presented them some major problems in an enterprise cloud implementation. The staff, which not only had to deal with this cloud roll-out but also recent staff reductions, was becoming overwhelmed trying to keep the systems performing at optimum levels. They needed some CLEVER ways to gain access to the information already available from their systems. They needed additional help in three areas: 1. Linux Guest 3 was unavailable and operations only knew about it when a trouble ticket was opened by an end user. Since they planned on adding additional Linux Guests they needed a way to proactively understand the Linux guest availability. The 'network team' had no responsibility for the 'guest LAN' and it was up to the mainframe support team to operate and support all the elements within the System z. 2. Additional Linux Guests were planned that would significantly increase traffic through the OSA adapter. In the past they didn't pay much attention to the OSA adapter because of over-provisioning for just their z/os traffic. Now that they were adding Linux traffic over this interface they wanted to understand the baseline for OSA transmissions and provide information to the capacity planning team. They also wanted to ensure that their z/os traffic was not impacted by all the Linux additions. They also were planning several Linux on System z implementations that would only have the z/vm operating system (no z/os operating system) and they needed to ensure that they could get OSA information in that environment. 3. Linux Guests 1 and 2 were primarily used as data backup systems for servers in the corporate network. The traffic was very predictable and occurred during 'off peak' hours. With the addition of Linux Guest 3 they were running production applications with interactive traffic requiring user acceptable response times and traffic flows not only out to the business network but also over to the z/os 2
databases. How could they tell if traffic was using the HiperSockets connection? Problem 1 - Availability Monitors like CleverView for TCP/IP and CleverView for TCP/IP on Linux provide many functions but one of the most basic functions is the ability to understand availability. In this case using these monitors allowed operations to understand in real time the status of their IP resources; critical servers, network connections, the critical applications running in each server, and the important links (OSA adapter, HiperSockets LAN links, and Guest LAN links). Using the alerting function operations would be alerted of any state change to the defined and monitored resources. Note that each resource's alerting function was individually set based on the criticality of the resource and its impact on the overall business operation. These availability alerts were not only sent to the CleverView Operations Interface screen, but also to their business event console which could receive the SNMP traps that the CleverView products produced. Note that in reality alerting was set on many other resource elements like CPU usage on Linux Guest 3. Problem 2 - OSA Adapter Monitoring I have worked with so many of you who have indicated that you don't monitor your OSA adapter because it works and your z/os environments are mature. Like the client in this scenario, this attitude may and should change once you add Linux Guests to the workload mix. In this case we are adding three new servers and all their workloads are utilizing the OSA adapter as the main connection point to the business network. With Linux Guest 1 and 2 the traffic was very predictable in both amount and execution time. Linux Guest 3 now added interactive Linux application traffic which meant new insight was needed into the OSA adapter. The first step was to look at the OSA information they had been collecting with the CleverView monitor. Luckily they had set up OSA monitoring but just never looked at the data because it was working as required. They had over 2 years of historical data so we were able to get a good view of the baseline and determine if there were any 'seasonal' changes in the baseline. In this case the first item was to establish the OSA baseline in three areas: OSA Express channel utilization OSA Express LPAR table utilization OSA Express Ethernet utilization Exporting this data into an excel spreadsheet allowed us to easily manipulate the data and come out with a clear view of their OSA utilization. One of the items that was discovered in this investigation is that on Tuesday of the last week of each month they had a 15% increase in traffic in non-peak hours. Further investigation showed that this was traffic going to the Linux Guest 1 and 2 systems. The operations team never realized that this was occurring until we started to look at the OSA information. Using the gathered information OSA thresholds were established to alert operations if utilization rose significantly outside the defined metrics. The challenge will be to continue to perform due diligence and re-inspect these metrics on a quarterly basis to ensure they are still valid. By the way this also prompted them to turn on monitoring of the HiperSocket and Guest LAN links so similar baseline information could be collected. They also analyzed the z/os TCP/IP monitored metrics to build another list of KPI (Key Performance Indicators) for their TCP/IP environment. Problem 3 - HiperSocket LAN 3
With the addition of Linux GUEST 3 they wanted to ensure that the Linux Guest 3 to z/os TCP/IP traffic was using the HiperSocket LAN and not going back out over the OSA adapter. The first step as noted above was to monitor the HiperSocket LAN traffic. Using near real-time (Snapshots in the CleverView products) and historical data views (Thru99 in the CleverView products) they were able to see the increase of traffic over the HiperSocket LAN. They were actually able to monitor both the z/os and the Linux sides of the HiperSocket LAN. As the operations team realized the extent of information provided by simple monitoring, they added the additional links to their monitored configuration and were able to find a congestion point in the Linux Router. The insight provided by simple monitoring allowed them to avert a potential performance incident. Upper management was very impressed with the amount of detail provided on the information flows and approved additional Linux migrations to the System z environment. AES Products CleverView for TCP/IP on Linux v2.2 and CLEVER Mobile TM for Linux Enterprise knowledge workers need access to server performance and availability details from not only their browser desktops, but also their cell phones, and they need a more concise way to view server status. CleverView for TCP/IP on Linux monitors and supports the complexity of a cloud environment, providing real-time, near-time, and historical data for such critical system resources as CPUs, ports, links, routers, HiperSockets, OSA and other devices. CleverView for TCP/IP on Linux enhancements address these issues and continues our unique support of both Linux on System z and open systems Linux. CLEVER Mobile for Linux empowers IT staff members to provide exceptional service to the business with their ios, Android and BlackBerry powered mobile devices. The added access capability ensures real-time notification of problems leading to increased Linux service availability, regardless of the hardware platform on which Linux resides - open systems or System z. Historical information access through these mobile devices improves management of service level objectives globally. Sample highlights: LinuxView is the focal point from which you can monitor the health of all Linux nodes and identify performance bottlenecks, providing alerting before users complain. LinkView monitors the network interfaces such as OSA Express and HiperSockets with traffic statistics and interface status. ProcessView provides the ability to interrogate application workload per process for CPU and memory. CleverView for TCP/IP on Linux is the solution that improves the agility, reliability, and performance of cloud - the foundation for your datacenter consolidation initiatives - by: Focusing on Consolidation and Virtualization Reaping the Most Business Benefits through Infrastructure Virtualization Assuring Application Service Levels 4
For more information visit our product page and download the data sheet and press release. CleverView for TCP/IP v8.4 Now in Managed Availability CleverView for TCP/IP helps performance analysts, operations personnel, knowledge workers, system programmers,and capacity planners effectively monitor performance, and plan for the future. Its superior performance monitoring makes it the ultimate choice for large IBM -hosted data centers undergoing wide scale Business Services transformation. By combining real-time monitoring and alerts to potential problems with in-depth historical reporting of activity and trends, CleverView for TCP/IP gives you a proactive, comprehensive, automated solution to z/os TCP/IP workload performance and availability management. Highlighted new features of CleveView for TCP/IP 8.4 are: SysPoint Dashboard presents summarized metrics for monitored z/os systems in a graphical format. CICS Connections Monitoring reports on CICS connections on a transaction basis. Restructured Navigation Menu consolidates information into concise categories based on users' roles. Expert Reports zoom-in offers more efficient access to performance data allowing a single click on a date or hour field in Expert Reports to provide the corresponding detailed report. StackView Monitoring now provides historical metrics allowing trending analysis for each monitored address space. BatchPR adds a new TN3270 Session Performance Report allowing automatic report creation. For more information visit our product page and download the data sheet and press release. Need More Information? If you would like more information on CLEVER Solutions, please visit our Website. If you are interested in setting up an interactive Webinar, or would like to schedule a free 30-day trial, click on the highlighted area to fill out a web based request form, send us an email, or call us at (650) 617-2400 or (650)- 617-2401. 5
To ensure you receive future editions of the AES Newsletter, simply add news@aesclever.com to your email address book. CleverView, CLEVER, CLEVER TCP/IP, CLEVER eroute, CLEVER ctrace, CLEVER Buffer, CLEVER Web, CLEVER/SNA and CLEVER eperformance are registered trademarks of Applied Expert Systems, Inc. CLEVER Mobile is a trademark of Applied Expert Systems, Inc. The IBM logo, Business Partner emblem, zenterprise, z/os, and z/vm are trademarks of International Business Machines Corporation in the United States, other countries, or both. The HP Business Partner logo is a trademark of Hewlett-Packard Development Company, L.P. The Red Hat Ready ISV Partner logo is a trademark of Red Hat, Inc. in the U.S. and other countries. Used under license. The Novell PartnerNet Silver Partner logo is a trademark of Novell, Inc. in the U.S. and other countries. Microsoft and the Microsoft Partner Network logo are either registered trademarks or trademarks of Microsoft Corporation in the United States and/or other countries. Android is a trademark of Google Inc. BlackBerry, RIM, Research In Motion, SureType, SurePress and related trademarks, names and logos are the property of Research In Motion Limited and are registered and/or used in the U.S. and countries around the world. Used under license from Research In Motion Limited. ios is a trademark or registered trademark of Cisco in the U.S. and other countries and is used by Apple under license. All other trademarks are the property of their respective owners. Forward email This email was sent to laurak@aesclever.com by news@aesclever.com Update Profile/Email Address Instant removal with SafeUnsubscribe Privacy Policy. AES 149 Commonwealth Drive Menlo Park CA 94025 6