Systems Perspective on Software Development Project Cycle

The “Cloud” has become the dominant global trend in Enterprise IT today and the big question for many businesses is how to transform their IT model or adapt cloud initiatives to align with their strategy, structure and processes. I would like to take a step back and take a systems approach to analyze the challenges in the current software development life-cycle. I think a systems model will help in learning about the dynamic complexity of IT development projects and understand the sources of resistance to design more effective system models/policies for cloud transformation strategies.

I created a causal loop diagram to understand the challenges and relationships between different variables of the IT project cycle. The intent is to include sufficient number of key variables and their relationships to represent reality.

The Causal Loop Diagram shows four important loops:
   1. Product Development (Balancing loop)
   2. Product Management (Reinforcing loop)
   3. Employee Productivity (Balancing loop)
   4. Product Maintenance (Reinforcing loop)

The balancing Product Development loop illustrates that when there is an increase in customer requirements, need for scoping the existing software functionality and architectural design decisions increase as well. The reinforcing Product Management loop suggests that perturbations such as management actions can lead quickly to major changes in project schedule, prioritization and performance.

There is one key negative feedback loop - Employee Productivity. This loop reflects the assumption that if a task falls behind schedule or there is a critical customer requirement, management will direct workers to work overtime to put the task back on schedule in order to satisfy customer deadlines. The Product Maintenance loop represents the post development maintenance activities that involve deploying the software to meet customer business requirements.


Among other things, I think one of the significant advantages for software development teams will be to leverage the cloud vendor for the infrastructural related maintenance activities. Another causal loop diagram depicting the cloud IaaS/PaaS/SaaS model will reveal the finer details. Will get to it next ...

Linking Congruence Model of Change with Balanced Scorecard and BPM

    I recently came across David Nadler’s theory (1995) - Congruence model of Change. The proposed model is an open systems model based on the proposition that the effectiveness of an organization is determined by the congruence between the various elements of the organization. Congruence is "the degree to which the needs, demands, goals, objectives, and/or structures of one component are consistent with those of the other".

    In analyzing the applicability of the congruence model, the inputs towards say, for example IT-Cloud transformation initiative for an organization, include the external environment, internal resources (e.g. people, knowledge skills, technology stack) and the organizations’ history. Based on these inputs, the executive management formulates the strategy for initiating changes in the IT delivery model. The outputs are the performance of the various sectors of the organization after the changes are implemented. I took a systems approach to depict the broad considerations and tools that can be utilized to drive the transformation processes.

1. INPUTS – Causal loop diagrams can be a good choice here to truly understand the organizational environment and the dynamic system by observing variables and their cause of variations.
2. Strategy - Kaplan & Norton' performance measurement framework ‘Balanced Scorecard’ can be used to align business activities to the vision and strategy of the organization, improve internal and external communications, and monitor organization performance against strategic goals.
3. Process Perspective – Business Process Management is a coherent and consistent way of understanding, modeling, analyzing, executing, monitoring and optimizing business processes as well as associated resources leading to business improvement. BPMN tools offered by Oracle SOA Suite can be employed here.
4. OUTPUTS – Business Intelligence and enterprise performance management tools deliver insights to help managers make better-informed decisions and are conducive to the enhancement of key performance indicators (KPIs).

Oracle ESS SubRequests

Oracle Enterprise Scheduler Services (ESS) provides a powerful and useful mechanism to process data in parallel - "Subrequests". The common use case is to employ ESS Subrequest feature for parallel execution of child jobs. For example, parallel processing of payroll for a large number of employees can be achieved by executing subjobs or child jobs i.e one job for each letter of the alphabet.  This feature combined with ESS support for handling job dependencies (and incompatibilities - a ESS feature) will ensure that a payroll job does not run at the same time as a salary increase job.

In simple terms, an ESS job request is considered as a subrequest when a running job submits a new request, passing its own execution context. Passing the execution context ties the request being submitted to the currently running request.

Once a parent request submits a SubRequest, that parent must return control back to Oracle Enterprise Scheduler, in the manner appropriate for its job type, indicating that it has paused execution. Oracle Enterprise Scheduler then sets the parent state to PAUSED and starts processing the sub-request. Once the sub-request finishes, Oracle Enterprise Scheduler places the parent request on the ready queue, where it remains PAUSED, until it is picked up by an appropriate request processor. The parent is then set to RUNNING state and re-run as a resumed request.

In general, SubRequests are treated like any other request in regard to their processing. ESS will not automatically update the parent request status to ERROR on failure of a sub-request. It is up to the parent job to determine its own tolerance for sub-request failures and complete with the appropriate request status based on the outcome of the sub-requests.

Some important points to remember:

• ESS does not keep any kind of "ordering" in regard to subrequests submitted by a given execution of the parent request. Once the parent pauses, the processing of subrequests is like any other request; i.e., it is subject to workassignments, incompatibilities, async throttling, available worker threads, and so on
• There is no internal restriction imposed by ESS that limits how many sub-requests could be executing at a given time
• ESS WebService does not support sub-request submission 
• Cross-cluster sub-requests are not possible

Please refer the Oracle ESS Developers Guide for more details/latest information.

Monitoring application performance and stuff you can do with NetBeans and VisualVM

    "Patience is a virtue" – probably not applicable when it comes to working with enterprise web applications, where the average web/mobile user expects pages to load in two seconds or less.  The sheer complexity of IT environments handling high transaction volumes and complexity is the norm, and without application performance monitoring, it will be impossible to deliver consistent service levels.

    As enterprise Java technology becomes more and more pervasive, understanding the behavior of the garbage collector becomes extremely important to monitor/manage application performance. Garbage Collection (GC) involves traversing Java heap spaces where application objects are allocated. So, it is important to know how frequently garbage collections are occurring and the time it takes for JVM to free memory no longer utilized by application logic. Heavy or excessive garbage collection will often show up as high CPU (the threads identified in high CPU will be the GC Threads).

    The amount of time that garbage collection takes is related to the number of live objects in the heap space and not necessarily related to the size of a given Java heap size. The popular in-built or open source tools one can leverage to optimize the JVM for throughput and responsiveness are VisualVM (included with JDK 6 update 7 and later) and NetBeans Profiler.

Here is a list of performance monitoring/profiling stuff that can be performed using VisualVM and NetBeans:

1. Capture sophisticated heap profiles using NetBeans profiler. Heap profiling is beneficial for observing frequent garbage collection and provides information about the memory allocation footprint of an application
2. NetBeans CPU Profiling using bytecode instrumentation to profile the performance of an application
3. Memory Profiling using VisualVM to analyze the memory usage of an application 
4. View live heap profiling results and traverse object references using NetBeans HeapWalker
5. Visually observe garbage collector behavior using VisualVM plugin "VisualGC"
6. Monitor surviving generations (or object age); an increasing number over a period of time can be a strong indicator of a memory leak 
7. Monitor Class loading and JIT compilation using VisualGC 
8. The number of threads that are currently active within your application, along with a drilldown of what type of threads that are in use 
9. VisualVM displays real-time, high-level data on thread activity, and thread dumps are very convenient to help diagnose a number of issues such as deadlocks or application hangs
10. Analyze code fragment performance and identifying CPU bottlenecks using NetBeans

Why Distributed Caching Matters ...

     My online shopping experience is limited to electronics and books, and the web site that provides good, fast and secure user experience invariably becomes my vendor of choice. For many web applications, consumers are likely to access the same data many, many times over the course of a day, and they all want fast concurrent access to, say item prices and other information. Behind the scenes, web and grid computing applications may repeatedly retrieve popular data, such as product descriptions, or they may create application-wide data, such as game scores, schedules, or interim stock trading results, all of which have to be accessible to all application servers deployed in a data grid. So, from a business perspective, it is important to recognize that there is an adverse impact on revenue when web page loads slowly and access to information is frustrated.

     Challenges: A rising necessity to share more users’ session data between different Web applications, domains or application servers is a major challenge in designing distributed architecture. With the decline in exponential growth of CPU speed over the past few years, the onus is on the developers to wring out every microsecond of latency and to maximize application throughput. So, the paramount need for deploying high-performant web environments, is to move away from the methodology of having applications query the database or metadata store directly each time data is required to be retrieved, updated or passed around, to one where data lives closer to the application tier. This is where data grid solutions come in, where its innovations in distributed caching make their mark.

     Business Solution:  The main goal of Distributed Caching (or Data Grid) is to provide as much data as possible from memory on every grid node and to ensure data coherency (distributed transactions). By reliably staging application data in memory within the compute grid's memory, data is simultaneously available to all compute nodes with very low latency. This facilitates distributing application’s workloads and the result is many fewer round trips to the database and faster access to data at in-memory speeds.

     Milliseconds translate to significant dollars spent or saved and distributed caching reduces not only the cost but also the risk as compared to scaling up lower data-tiers.For web applications, the in-memory data grid solution enables organizations to economically address data demand usecases involving repetitive reads/writes and session state management. In the era of cloud computing, caching (access to in-memory data) plays a pivotal role in the design of distributed systems.

      Data Grid/Cache Vendors:  Both commercial and open source offerings are available. JBoss Cache, GridGain and EhCache are some of the popular Java open source software for processing in-memory data. On the commercial side, for example, Oracle Coherence lets session states be managed in a variety of caching topologies and enables session data to be stored outside of Java EE application servers. This means that application server heap space is freed up and servers can restart without session data loss.

     Browsing through the product offerings will generally give you an idea of the advantages and solutions readily available for leveraging important application performance gain, and why distributed caching matters.

A look at Android Architecture

Google’ Android documentation (http://developer.android.com/guide/basics/what-is-android.html) has come along pretty good. I now see that there is a cool Android Architecture diagram that shows the major components of the Android OS.

Summarizing the 10 notable aspects of the architecture that every developer should be aware of:

  1. Android is built on portable Linux platform that brings a level of hardware abstraction

  2. Android applications run as separate Linux processes with security delegated to the underlying Linux system
  3. The native libraries are C/C++ libraries and provides necessary services to the Android application layer

  4. Bionic standard C library is rewritten to make it license friendly and purpose-built for tiny, battery-powered devices

  5. Includes Apache Harmony, a modular Java runtime with class libraries

  6. Relies on Linux core kernel for memory management, power management, and networking features

  7. Java VM was replaced with the Dalvik Virtual Machine to focus strictly on mobile devices and avoid licensing

  8.  Dalvik VM executes files in the Dalvik Executable (.dex) format which is optimized for minimal memory footprint

  9.  Includes SQLite, a full featured relational database engine available to all applications

  10.  3D graphics is based on the OpenGL ES 1.0 specification

Android SDKs - Cupcake to Ice Cream Sandwich

My journey with mobile’ leading development platform, Android started back in early 2011 with my first Android app (weekend project) in the marketplace - ‘tennis grand-slam champions’. Ever since, I am keeping a tab on the Android platform and following its feature offerings as time permits.

In the mobile development world, the release cycles are considerably shorter than the traditional enterprise application development. Android’ first stable version ‘Cupcake’ was released in April, 2009 and within couple of years; we have seen many features evolved in ‘Ice Cream Sandwich – Android SDK 4’. Éclair (v2) was better optimized than Donut (v1.6) to take advantage of the hardware for faster processing. Speed improvement was brought in with the integration of Chrome V8 JavaScript engine and JIT optimization in Froyo (v2.2); bluetooth and wi-fi made its way into the networking layer.

Gingerbread (v2.3) was a major release and a considerable number of new features were introduced, that include new UI themes, redesigned keyboards, new copy and paste functionality, improved power management, NFC (Near Field Communication), support for VoIP/SIP calls, new Camera application for accessing multiple cameras and supports extra-large screens. Ice Cream Sandwich (v4) is much sleeker than Gingerbread and Google has designed a font completely from the ground up, which looks exceptional. Another major change is that a lot of the SQL handling is moved from the native to the Java layer.  I am sure there are lots of articles on the web that will give a quick comparison of the feature-sets that were bundled in every SDK release.

           iOS leads the way with quality apps, Android leads in market share; Windows phone has not caught up yet with the mass adoption as its leading competitors. Android and iOS — each offers its own clear take on the mobile OS experience, and I personally feel that the consumers and business users stand to benefit from this healthy competition. Staying tuned on what Android comes up with next …………………