Performance Engineering Management or PEPM

A "what is all the fuss about" discussion about the topic of Strategy and Performance Engineering;

- with Maxwell Mattison/Performance Architecture Engineering Consultant

Performance Engineering, Management...Project, PEPMAN[5]

Information Technology Infrastructure Library dictates there is a simple structure that IT organizations can follow. The structure allows each important role to be acknowledged simply by looking up the role of interest in Wikipedia or Wiki. There are also other institutions that are as highly regarded that readily compile this same information.

An interesting question asked...What does an IT Performance Engineering Management, or Performance Engineering Architect , or a PEM do? Reading through each or any description on the internet may not be much help.

This role/designation/skill-set is not to be confused with any person whom is already working as a Performance resource whom maybe already be involved in the Software Development Project. Or a Performance Tester whom maybe asked to run a series of scenarios and hand-off the test results to an analysis team for review kinda job requirement.

It is understood a PEM is a lot like a CSI Agent at times! One might ask what does a CSI Agent or Continual Service Improvement (CSI) Agent job description even consist (of)? Often times a Performance Engineering Management role is first aligning then realigning IT services with the Enterprise Architecture allowance or change as the business needs decline , and most often with software development products increase. CSI needs to be treated just like any other service practice.[1] A Performance Engineering Management resource typically gets involved with this initiative. With a little help, this role assists the Systems Architecture Group carrying out special Production Fire-storm scenarios needing resolved ASAP. The careful coordination is juggling a long list of Technology Stack Integration Points configured to be swapped out with higher performing technologies as the become tested and available for high endurance use.

The Performance Engineering Management resource keeps a head in the cloud at all times! It shouldn't matter if it's ORACLE Cloud, Amazon, and Azure Cloud (Microsoft), MicroServices (DXC), Private Data Center/Private Cloud Services the top few one can look forward to working with as they are polite with Vendor Management and even provide open-source or already purchased products for the team to discover. This is a rapidly changing or hyper-dynamic environment(s) as laptops and clunky PC are dumped for fastest computers on earth's surface in Cloud Computing Space. The ability for a team to learn things quickly, or understand the impact what's changing, and then getting the big-data analytics in place are needed to uphold the client's end of the business end or SLA is an absolute must and needs to be performed for each vendor sometimes on a daily basis.

The specific initiative of Service Agreement Management or formally known activity Service Level Agreement Management is a concern due to the cost of supporting the vendor client relationship. Many times the Total Cost of Operation of a vendor's products or services are compounded with the maintenance agreements adding to a hefty price tag for organizations that consume IT resources as an integral business processes. Often times interacting with a vendor necessitates a purchase of software and hardware for this virtualization system to run on. If a software product can exist on a server with other software applications Capacity Management should be addressed. These software applications have specific technology requirements declaring it a better fit on specific physical hardware devices it can be demonstrated. An example is some software products can only exist as a singleton, meaning only one can exist. Other times having multiple instances of the same software application allow for more traffic and more users requests fulfilled.

These two ongoing initiatives combined with their projects which then spur off other software development activities. The Business Analyst Department[2] defines the end user or client wants or needs. This transcription process is documented in typical business language, and that can mean vague or ambiguous words or explanations of concepts of how software should operate. Possibly the most important information generated from this client Business Analyst interview is what duration of use is this system planning on being used on a 24hour clock. Sounds easy enough, the performance engineer takes this list and transcribes it using scientific algorithmic approaches into Non Functional Requirements. This Non Functional Requirements list is used during a Performance Inspection.

In general, functional requirements define what a system is supposed to do whereas non-functional requirements define how a system is supposed to be. A performance engineer instinctively runs a set of tests to fulfill these conditions or often called scenarios. Comparing and contrasting of how the test results help the performance engineer offer optimization opportunities to increase the overall responsiveness of this complicated and very expensive Technology Stack.

If these Non Functional Requirements are fulfilled by the Production Systems on a High Availability Architecture it offers an organization to withstand projected traffic patterns or project what may happen during other more exaggerated number of users or transactions. A classic scenario is checking e-mail around 12 or 6pm on the internet. Seems a little slower by most users accounting what the user's observed download times were for completing that use case. A Performance Management Engineer takes this in account and optimizes the configurations of the integration point settings along with the known Key Performance Indicators and to attain level described in the Non Functional Requirements documentation. The Availability Management initiative is a success when all parts of the system work as it was designed to do and behaves the way it was discovered to in the Production Environment.

Luckily there are Enterprise software products that can be used to help organize these initiatives that spur Software Development Projects that employ consultants that allow a company to thrive, grow. The prediction what of how the system may react during an unusual traffic patterns and future growth patterns are business opportunity scenarios. Internally, we cite these extremes as Critical Transaction Use Cases for the technology vendors or the companies. As they employ consultants dispatched to the client sites to support them under extreme cases of database crashing in OPS/DEV which should happen in a very controlled manner to study each of the pre-conditions of replicating this most likely suspects for poor performance scenario.

A PEM is like a CIA Agent, Communication Information Administration Agent. Typically involves communication protocols, or a deeply nested table of the right people and their preferred method or secured transmission communication line. This communication begins as the Project Kick-off does to keep Project Stakeholders in the comfortable level of communication of the PRE-Production Environments Performance Tuning or break-in. Not every teammate may want to be contacted each time the server CPU trigger's an alert, etc,.

A PEM may behave much like Research Science Methodologies would dictate. It's very common need to have Rapid Knowledge Transfer session(s), but what's not that common is asking the entire Migration & Implementation Project Team to have a collective brainstorming session. A typical discussion or project meeting maybe in order, but a better way to spend efforts is conducting an Architectural Design Diagrams Review. Especially, if things are highly complex and very complicated it's best to set up everyone on the team as a "Scientist" and ask questions like "What happens when we do ABC to XYZ?". This is the time to evaluate hardware and software components layouts, and their transaction pathways. What happens when we use Capacity Tool on APPS + Vendor_HardwareStorageServer + VenderName_CloudServices = Performance Tuning Configuration Settings YEAR#3 Capacity Planning Thresholds! It's best to have the Science Minds poke holes in it, to see if the the thing leaks...sure memory leaks are fine also!

Being a PEM is a lot like winning the Lotto! OK, not quite as much fun as that. Selecting an Automation Emulation Tool that can quickly and with as little maintenance as possible to emulate the business transactions with a highest number of Virtual Users on the planet can be a lot like winning the Lotto!

A PEM Team is better than a BANK! Test Data and the Stored Information whilst the SUT maybe be under way poses a very large security risk. Many smart teams take the gamble enter Human Resources using real data. As the law would mandate we are talking about using Managed Code to conduct Performance Tuning Activities. So create an Admin (level) Performance Engineer on OEM, and that's the extent of the interest of what is contained in any database record. There is no time for reviewing anything I mean for crying out load "we are talking terabytes and then petaflops of data here..."[3] when a System Under Test is put through it's paces and has Performance Inspection Testing requested for compliance standards. As funny as that joke is to some of us people of Earth we now sift through Petabytes worth of data. Why and how is that possible. At ORACLE HQ there is ORACLE LABS(oratories)[4]. One simply needs to pull up the ORACLE Laboratory Performance Test Organizing all the streams of information input, plotting the overlay of the Transaction Response Times, TPS against the Key Performance Indicator Monitoring reports. Creating a graphical display of these conditions by using tools or one can simply use EXCEL "fx" options.

The demonstration of the fulfillment of the Non Functional Requirements list using an algorithmic template type format is known as a NFX Report (NFR acronym).

A PEM is always known as an IT Firefighter, the every day activities create its share of problems and with a little practice can be solved quickly with little as possible downtime of the SUT, Systems Under Test. The Problem Management skill-set is learned, within the problem management domain, the performance engineering practices are focused on resolving the root cause of performance related problems. These typically involve system tuning, changing operating system or device parameters or even software configuration settings.

The ability to proactively manage Continual Service Improvement projects, Capacity Management projects, and Availability Management projects successfully is the "crux" of the matter or the "ethos" of this consulting position. A well versed PEM tool box consists of multiple tools that accomplish the same business need. Tool Vendors come and go. Tool versions go out of service, and end of life dictates that there is a continual zest for accumulating knowledge on the newest versions of these tool-sets and their platforms to remain competitive in the Performance Tool Sector Market.

Understanding this condition exists for organizations, and the demand in the IT sector job market has a new role for an enthusiastic, energetic, overachieving Performance Engineer with Architecture skills. Honestly, sometimes the only Agent available to help reduce the Total Cost of Operation, TCO.

"Sources Cited" during this conversation about Tools/Methods/CSI Process, and further reference to the following public sources of information.

[1] . https://wiki.en.it-processmaps.com/index.php/ITIL_CSI_-_Continual_Service_Improvement

[2]. https://www.homeworkmarket.com/files/macrounit5-pdf

[3]

[4] . www.oracle.com/labs/PeopleSoft HCM 9.1 FP2 HR Self Service Using Oracle DB on an Oracle Database Appliance-2196086.pdf

[5]. https://en.wikipedia.org/wiki/User:ORACLEPerformanceEngineer/sandbox