In application performance monitoring, thresholds are usually defined as fixed thresholds which are easy to define manually, but suffer from over generalization – not taking into account a variable’s (values) behavior changes over time. Fixed threshold breaches are a lagging indicator – identifying an extreme failure (like 100% CPU utilization) after it has already happened [...]

In my last post I suggested that behavior analysis is critical to cloud application performance management because it is the only way for an application owner to find performance issues early on – before they affect users. So let say you found out that certain areas of your application are behaving unusually compared to past performance, and the application [...]

Managing SLAs and performance in the public cloud is inherently different from managing performance in the data center or even from a private cloud. There are various way to describe what the cloud is (as I mentioned in my last post “The Future of Capacity Planning in the Cloud“) – but it clear from a technical [...]

Capacity planning is an annual ritual for many organizations (and culminates around now, with the proposal for next years IT infrastructure budget). Traditional capacity assumes that compute power and storage are expensive, finite, relatively unflexible resources so there needs to be a plan to ensure the lead time needed  to put those resources in place. But with the [...]

I have been reading some marketing hype around converging the management of mainframe and the cloud (e.g. Mike Vizard’s article on CA’s vision of the future data center).  Sorry, but I don’t buy it – not that it isn’t a critical issue for the future of data centers it is just that they are missing [...]

I read really good analysis of the APM market by Bernd Harzog (which I guess is now an official market since Gartner published a Magic Quandrant - though in reality the market has been around for at least 20 years now). I think he hit a number of issues right on the money. One piece he missed is [...]

Read-Write web came out with a really interesting infographic regarding virtualization and the web titled “What Your Admin Should Know About Cloud Management” thanks to Solarwinds. The main take-aways for me were: 1. Anyone using a cloud is probably using a private cloud (by a wide margin) 2. Performance bottlenecks, capacity planning and storage- i\o [...]

I saw an interesting federal government report on the IRS mainframe environment with the sexy title of “Mainframe Computer Performance Is Being Actively Monitored,  but Defined-Service Agreements and Software Licensing Can Be Improved“. It gives a glimpse inside monitoring and capacity planning for a large mainframe installation – including utilization and capacity graphs for a [...]

Cloud computing is high on everyone’s list of technology trends to watch for (though Gartner just  degraded it to the number 10 trend for 2012). I actually don’t think that the public cloud was actually a real trend in enterprise data centers – but private cloud computing (and maybe hybrid) is certainly an important trend [...]

I was reading Gartner’s magic quadrant (MQ) report on application performance monitoring (APM). It is the first MQ on APM – and it made me smile. APM is a new and growing area in distributed systems (there are over 25 vendors in the Gartner MQ) – even though APM has been around forever in mainframes. [...]

In my last past on “Application Performance Management – Prevention vs. Treatment” I wrote that predictive analytics doesn’t enable a “Minority Report” like capability. Then I happened upon this article “Cops on the trail of crimes that haven’t happened“, where predictive analytics are being used to help predict where crimes might happen in Santa Cruz. [...]

When you talk to organizations about application performance management most will tell you that they manage performance in two ways – in development through QA and testing (aka prevention) , and in production through tuning, monitoring and operations (aka treatment). The only person that cares about the whole picture is the application manager – but realistically the [...]

Thresholds are the standard mechanism used to define the bounds of expected (or normal) performance of mainframe system variables like CPU, dispatching wait time, DB buffer usage or the current number of active DB locks. Once thresholds are breached, the system is no longer acting normally – which can be indicative of an unusual production [...]

The role of APM (Application Performance Monitoring) technologies is to provide the data needed for application performance management processes and best practices (confusingly also known as APM). No matter whether you are using APM for mainframes or distributed systems, a key question is how do you know that application performance is improving as a result [...]

I was reading a blog post on “Service Performance and Availability Issues in the Cloud“, where the author talks about managing SLAs and the cloud. The good news from an SLA perspective is that clouds offer new options of unlimited dynamic resource allocation. If your applications are backing up – voila – click a few [...]

IT service levels are the critical component in the perceived performance of IT (as perceived by the business and customers). There are many service levels that need to be monitored (application development time, time to add new features, time to roll out an application), but none is more critical in today’s self-service, web enabled business world than [...]

If you look at software testing – performance testing is one of the hardest problems. There are load tests and stress tests, which use scripted assumptions about how a system will be used, and then use those scripts to load the system and see how it stands up. I think that most testers understand that this doesn’t [...]

I read an interesting article on the Noisy Neighbors problem in multitenant cloud infrastructure (which is where all public clouds need to be if they are to provide a true cost advantage over doing it yourself).  The article is based on Netflix’s experience with Amazon’s cloud. Because cloud infrastructure can put multiple tenants on the [...]

I was reading a short report about best practices around adding capacity to an existing mainframe system. The final analysis was that if you have a production mainframe with mixed on-line and batch workloads you should upgrade system capacity when your system utilization reaches 70%-85% of system capacity, the lower number (70%) for transaction only [...]

I had a conversation the other day about end-to-end application performance management. We were talking about diagnosing problems in a datacenter that has transactions that span both distributed systems and a mainframe. I wanted to talk about monitors, but they really wanted to talk about management. A typical techie vs. business view of the world. [...]

Alerts are the mechanism used by performance monitors to lower the information overload caused by the wealth of performance related data available for operators and technical. A good alerting system enables users not to have to watch the monitors at all times, its job is to inform them about when (and where) they need to [...]

Mainframe monitoring tools provide a wealth of information about the current state of your system. The problem is that there is just too much information available for a human to assimilate, so to make most of monitoring tools the best users have an implicit monitoring strategy they use to achieve the most value from their [...]

From my perspective, current cloud computing infrastructure is all about abstracting away hardware from the software infrastructure – or in other words hardware virtualization. Software executes on virtual machines that reside on physical machines – but the software doesn’t really care at all about the physical machines, all it cares about are the attributes of [...]

Combining state-of-the-art behavioral performance analytics and application performance tuning tools enables a highly performance tuned mainframe environment.  A mainframe isn’t a static environment – but rather a living environment with a complex lifecycle. There are workload changes, software and hardware updates, data and file modifications, user profile changes and new applications changing the mainframes environment [...]

“As we go from customer to customer, it is amazing to see the rapid acceptance of the ConicIT concept and the willingness of companies to do a POC.  It is so clear that the mainframe remains crucial to the large financial institutions, as well as other organizations, and their is a real lack of applications [...]

Today’s large datacenters are hybrids – they manage both distributed systems (e.g. Linux) and mainframe systems.  Many applications span both – perhaps starting as a transaction from a web customer, but then fetching relevant customer information from the mainframe. From a customer’s perspective it is a single seamless system, the problem is that technically (and [...]

Mainframes are very stable, high capacity systems. However, even the best tested and best run mainframe systems have production performance problems. One of the key reasons (besides human error) is the high degree of virtualization in mainframe systems, where essentially all resources are virtualized. That level of virtualization causes complex (and unexpected) interdependencies and interactions [...]

We will be at Share 2011 in Orlando at the SDS booth at the SHARE Technology Exchange Expo, Booth # 201 We will also be presenting at the conference, giving a general overview predictive analytics and of how it can be used in a mainframe prouction environment, to prevent recurring performance problems and lower costs. If you are [...]

Hi everyone. This is our first post. We are a software vendor specializing in the development of a unique mainframe behavioral performance analysis solution for first fault problem resolution. The company’s goal is to employ its sophisticated mathematical tools to help enterprises improve their service levels, prevent (recurring) service problems, and reduce costs. This post [...]