During the past 7 business days I took part in a training about SIMATIC IT XHQ 4.0 software.
I participated in the Basic (21-22 Jan) and Advanced (25-29 Jan) training.

This training was given by the engineer Fabio Terasaka who’s a team lead at Chemtech having over 3 years of experience using and deploying XHQ in Brazil and internationally.

I decided to write this post so that people can get an overview of XHQ from a consultant/developer perspective.

I’m also excited about the endless possibilities XHQ has to offer when it comes to optimizing and applying intelligence to an enterprise.

What is XHQ?
For those who don’t know or who have never heard of XHQ, here goes a succinct description of it extracted from its official site:

SIMATIC IT XHQ Operations Intelligence product line aggregates, relates and presents operational and business data in real-time to improve enterprise performance. Through SIMATIC IT XHQ, you have a single coherent view of information, enabling a variety of solutions in real-time performance management and decision support. ^[1]

XQH extracts data from a variety of systems - as the production (PIMS), laboratory (LIMS) and plant-floor systems. XHQ unifies all the operational and management data in a single view, in real time, allowing you to take a snapshot, minute to minute or second to second, of all the enterprise.

XHQ can be integrated in the intranet or a website for operations management, integrating production data such as the use of raw material and equipment, stocks, as well as data related to the product (temperature, pressure, electrical current), quality and maintenance.

XHQ implements the concepts of Operational Dashboard and Management ^[2] by Key Performance Indicators (KPIs) ^[3].

XHQ is used in energy, petrochemical, and manufacturing industries to aggregate, draw relationships, and then graphically depict business and operational data.

XHQ Timeline
XHQ was created in 1996 by an American company called IndX Software Corp based on Aliso Viejo, California, USA.

In December 2003 Siemens expands its IT portfolio acquiring IndX ^[4].

In December 2009 Chemtech - A Siemens Company absorbs the company responsible for XHQ around the world ^[5].

XHQ Architecture
XHQ has a modular architecture as can be seen in the following picture:

XHQ Architecture Overview

Back-end Operational Systems
Comprised of databases and its respective connectors that give access to business real-time data: times series data (PHD, PI, OPC), real-time point data (Tags), relational databases (Oracle, MS-SQL), enterprise applications (SAP), etc.

Middle Tier
Comprised of XHQ set of servers. Each XHQ server plays a role in the system:

XHQ Enterprise Server manages the end-users views of data that are created by XHQ developers.

XHQ Solution Server has the Real-time Data Cache and the Relational Data Cache that removes the burden associated with backend data retrieval.

XHQ Alert Notification Server (XANS) is a subsystem of XHQ responsible for alerting end-users about any inconsistence existent in the system.

3rd Party Web Servers as IIS and Tomcat give end-users access to data processed by XHQ.

User Interface
Users can access XHQ processed data (Views) using PDAs, web browsers, etc.

Users also have access to View Statistics that is a kind of Google Analytics. It shows default reports about peak and average user count, user and view hits by month, user and view hits by week, view usage by user per day, view usage per day, etc. You can create your own analytics reports using custom SQL.

Starting with XHQ 4.0 there’s a separate application called Visual Composer that enables developers to create dynamic, high customizable data views. Visual Composer can use XHQ data collections as its data source.
Visual Composer focus in graphics/charts and tables/grids to show business strategic content.

XHQ behind the Curtains
XHQ does its job using a subscription model based on the client-server architecture. Clients are automatically notified of changes that occur on the monitored variables. For example, if a user uses a view that has 2 plant variables and their pool period (configured in the connector or on the variable itself) is set with 2 seconds, the user screen will automatically refresh (using Ajax) to show the new variable values at each 2 seconds. This is the so called real-time process management.

XHQ core is implemented in Java and uses a Java Applet that is loaded in the browser to present the data to the user.

XHQ makes extensive use of JavaScript to inject customization points into the software.

Servers configuration are kept in .properties files making it easy to edit.

Data presented to the user comes from “Collections” that use high performance data caches that are XHQ own local databases. You can use live data from the backend but it’s not advisable because of the overhead implied. The performance gains can be better verified when lots of users are using the same view.

Skills demanded by XHQ
To get XHQ up and running you’ll need the following skills:

SQL query skills. SQL is used all the time to retrieve the right data from the back-ends.

XML and XSLT skills. Both necessary to configure data points (Tags) in the system and to export data.

Previous software development skills using the .NET Framework or Java are important to develop extension points to XHQ.

JavaScript skills. Used to define custom system variables and client configuration.

HTML and CSS skills. Used to customize the user UI.

Web server administration using IIS and Tomcat is a plus when it comes to deploying the solution in the customer.

Computer network skills. Used to detected any problem between clients and servers.

Solid debugging skills related to the above mentioned technologies. If something goes wrong, you’ll need to check a lot of log files (there is one for each agent in the system).

XHQ Implementation
XHQ consultants/engineers are the guys responsible for studying the necessities of the customer interested in optimizing the enterprise.

The following are 10 basic steps used when XHQ is implemented as the choice for business optimization:

01 - XHQ is installed on client premises;
02 - Groups of users and use cases are defined;
03 - Connectors are created to access data sources scattered all over the enterprise;
04 - A solution model is defined;
05 - A navigation model is defined; 
06 - Views of data for different audiences and activities are built;
07 - System components and collections of data are linked for data retrieval;
08 - The solution is updated, tested and optimized;
09 - Steps 1 through 7 are iterated;
10 - Security is applied in the solution model through the use of roles/user groups.

XHQ Value as a RtPM Tool
XHQ aggregates value to your business as a RtPM (Real-time Process Management) tool:

Using XHQ, operational costs may decrease an average of 8% each year, while the production of high value products may increase 10.5%. This is because XHQ helps the management board in the decision taking process. ^[6]

The following are 10 basic reasons why XHQ aggregates value to the business:

01 - Directors and staff can take their decisions based on the same information;
02 - Response times are dramatically reduced;
03 - Information availability is made true from one area to another (and vice-versa);
04 - Interfaces between one area and another, usually managed by different teams and systems, can be closely monitored;
05 - User-friendly and self-explained process schematics simplify plan management;
06 - Reduced load on mission-critical systems: read only users can use only XHQ;
07 - Leverage of other investments: PIMS systems utilization is increased, and become mission-critical as well;
08 - Intangible gain: re-think strategies for company needs in terms of what information is
considered critical for business decisions;
09 - Integration with enterprise applications as SAP R/3, logistics is greatly improved by
watching supply and distribution movements;
10 - Transport can come and go (monitored) graphically.

XHQ Customers
XHQ is used throughout the world.

The following are some of the customers already using XHQ to optimize their business:

CSN, ExxonMobil, Chevron, Dow Chemical, Saudi Aramco among others.

Interested in optimizing your business?
If you’re looking for business optimization/intelligence, you can get in contact with Chemtech for more information.

References
[1] SIMATIC IT XHQ official site

[2] Business Performance Management

[3] Key Performance Indicator

[4] Siemens expands its IT portfolio in process industries (PDF file)

[5] Chemtech absorbs the company responsible for XHQ around the world

[6] Siemens of Brazil Press Information (in Portuguese)

[7] XHQ for Steel Mills Real Time Performance Management (PDF file)

[8] XHQ can gather information from the whole oil & gas production chain

[9] Chemtech enters into the IndX’s biggest XHQ project in Brazil

I had to install the Blogs portlet in Liferay.

Liferay is the all purpose portal framework that Chemtech uses to build its website.

The Liferay portal already deployed on production server is the 3.4.5 version. When I tried to add the Blogs portlet through the Add Content menu option I couldn’t find it.

Googling about Liferay’s Blogs portlet didn’t help me. The only positive clue I had was

Liferay Portal Administrator's Guide, Third Edition

(page 124) which has a section dedicated to the Blogs portlet.

I tried to understand why the Blogs portlet wasn’t available in the Add Content window:

Was it because the blogs portlet didn’t make it into the version 4.3.5 of the portal? The answer is no. The blogs portlet is available in version 4.3.5 (with limitations if compared to the Blogs portlet of today’s Liferay version that is currently 5.2.3).

After a little bit of more googling I found Development in the ext environment wiki article. I read in item 4 that you can turn portlets you want to deploy on/off by editing the file

\ext\ext-web\docroot\WEB-INF\liferay-portlet-ext.xml

Mine was located in

E:\chemsite\tomcat\webapps\lportal\WEB-INF\liferay-portlet-ext.xml

I did just that turning the Blogs portlet ON setting the <include> property to true:

<!--
    Liferay Portlets

    To create a minimal installation of Liferay so that only the essential Liferay portlets are available, uncomment the following block and set the include attribute to false for the portlets you want to remove. To make a portlet available, set the include attribute to true. The struts-path attribute is shown so that it's easier for the editor of this file to associate a portlet id with a portlet.
-->

<portlet>
          <portlet-name>33</portlet-name>
          <struts-path>blogs</struts-path>
          <include>true</include>
</portlet>

I then rerun Liferay portal using Eclipse. For my surprise I could find the Collaboration category in the Add Content window with the Blogs entry available:

Hope this shortens the path when you come to need to turn a portlet on/off.

A* pathfinding search in C# - Part 1
A* pathfinding search in C# - Part 2

Code available at GitHub: https://github.com/leniel/AStar

This is the last installment in the series about A* (A star) search.

The C# source code implemented is available in the final part of this post.

As promised in the last words of A* pathfinding search in C# - Part 2 today we’re gonna run a test case using the Romania map.

If you want to understand the whole process implemented in this solution, please start reading A* pathfinding search in C# - Part 1.

When you run the console application, you get the following screen:

You start by entering a Start and a Destination city picking up the ones you want from the list of Romania cities.

When you press Enter the console app will show you the shortest or best path based on the A* search algorithm.

As you can see in the above screenshot, the app shows us that the best path to go from Arad to Bucharest is the one that goes as follows:

From Arad           to  Sibiu          -> Total cost = 223.236 km From Sibiu          to  Rimnicu Vilcea -> Total cost = 301.317 km From Rimnicu Vilcea to  Pitesti        -> Total cost = 348.536 km From Pitesti        to  Bucharest      -> Total cost = 456.108 km

Note that the Total cost is the cost calculated so far for each path, that is, in the example shown above, Total cost = 348.536 km is the distance in kilometers for travelling from Arad to Pitesti.

No doubt this is the shortest path to follow if you plan to go from Arad to Bucharest. We could choose different possible routes but the total distance traveled would be greater than the one the app calculated for the shortest path. Let’s see why this is so using the method ViewOtherPaths (I commented about it in A* pathfinding search in C# - Part 2).

The following is the output of the console app when the method ViewOtherPaths is uncommented inside the FindPath method. This helps you debug and see why the app has chosen the above shortest path.

A* Search - Sample implementation by Leniel Macaferi, June 7-20, 2009

These are the Cities you can choose as Start and Destination in Romania:

Arad
Bucharest
Craiova
Dobreta
Eforie
Fagaras
Giurgiu
Hirsova
Iasi
Lugoj
Mehadia
Neamt
Oradea
Pitesti
Rimnicu Vilcea
Sibiu
Timisoara
Urziceni
Vaslui
Zerind

Enter a Start city: Arad

Enter a Destination city: Bucharest

Possible paths:

From Arad           to Sibiu          -> Total cost = 223.236 km
Estimation          = 213.803 km
Priority Queue Cost = 437.039 km = (Total cost + Estimation)

From Arad           to Timisoara      -> Total cost = 48.459 km
Estimation          = 408.79 km
Priority Queue Cost = 457.249 km = (Total cost + Estimation)

From Arad           to Zerind         -> Total cost = 51.908 km
Estimation          = 431.034 km
Priority Queue Cost = 482.942 km = (Total cost + Estimation)

Possible paths:

From Arad           to Sibiu          -> Total cost = 223.236 km
From Sibiu          to Rimnicu Vilcea -> Total cost = 301.317 km
Estimation          = 154.102 km
Priority Queue Cost = 455.419 km = (Total cost + Estimation)

From Arad           to Timisoara      -> Total cost = 48.459 km
Estimation          = 408.79 km
Priority Queue Cost = 457.249 km = (Total cost + Estimation)

From Arad           to Sibiu          -> Total cost = 223.236 km
From Sibiu          to Fagaras        -> Total cost = 287.59 km
Estimation          = 178.296 km
Priority Queue Cost = 465.886 km = (Total cost + Estimation)

From Arad           to Zerind         -> Total cost = 51.908 km
Estimation          = 431.034 km
Priority Queue Cost = 482.942 km = (Total cost + Estimation)

From Arad           to Sibiu          -> Total cost = 223.236 km
From Sibiu          to Lugoj          -> Total cost = 397.029 km
Estimation          = 356.126 km
Priority Queue Cost = 753.155 km = (Total cost + Estimation)

From Arad           to Sibiu          -> Total cost = 223.236 km
From Sibiu          to Arad           -> Total cost = 446.473 km
Estimation          = 420.536 km
Priority Queue Cost = 867.009 km = (Total cost + Estimation)

From Arad           to Sibiu          -> Total cost = 223.236 km
From Sibiu          to Oradea         -> Total cost = 444.358 km
Estimation          = 434.745 km
Priority Queue Cost = 879.104 km = (Total cost + Estimation)

Possible paths:

From Arad           to Sibiu          -> Total cost = 223.236 km
From Sibiu          to Rimnicu Vilcea -> Total cost = 301.317 km
From Rimnicu Vilcea to Pitesti        -> Total cost = 348.536 km
Estimation          = 107.572 km
Priority Queue Cost = 456.108 km = (Total cost + Estimation)

From Arad           to Timisoara      -> Total cost = 48.459 km
Estimation          = 408.79 km
Priority Queue Cost = 457.249 km = (Total cost + Estimation)

From Arad           to Sibiu          -> Total cost = 223.236 km
From Sibiu          to Fagaras        -> Total cost = 287.59 km
Estimation          = 178.296 km
Priority Queue Cost = 465.886 km = (Total cost + Estimation)

From Arad           to Zerind         -> Total cost = 51.908 km
Estimation          = 431.034 km
Priority Queue Cost = 482.942 km = (Total cost + Estimation)

From Arad           to Sibiu          -> Total cost = 223.236 km
From Sibiu          to Rimnicu Vilcea -> Total cost = 301.317 km
From Rimnicu Vilcea to Craiova        -> Total cost = 400.614 km
Estimation          = 183.042 km
Priority Queue Cost = 583.656 km = (Total cost + Estimation)

From Arad           to Sibiu          -> Total cost = 223.236 km
From Sibiu          to Rimnicu Vilcea -> Total cost = 301.317 km
From Rimnicu Vilcea to Sibiu          -> Total cost = 379.398 km
Estimation          = 213.803 km
Priority Queue Cost = 593.201 km = (Total cost + Estimation)

From Arad           to Sibiu          -> Total cost = 223.236 km
From Sibiu          to Lugoj          -> Total cost = 397.029 km
Estimation          = 356.126 km
Priority Queue Cost = 753.155 km = (Total cost + Estimation)

From Arad           to Sibiu          -> Total cost = 223.236 km
From Sibiu          to Rimnicu Vilcea -> Total cost = 301.317 km
From Rimnicu Vilcea to Mehadia        -> Total cost = 461.891 km
Estimation          = 299.853 km
Priority Queue Cost = 761.744 km = (Total cost + Estimation)

From Arad           to Sibiu          -> Total cost = 223.236 km
From Sibiu          to Rimnicu Vilcea -> Total cost = 301.317 km
From Rimnicu Vilcea to Lugoj          -> Total cost = 504.328 km
Estimation          = 356.126 km
Priority Queue Cost = 860.454 km = (Total cost + Estimation)

From Arad           to Sibiu          -> Total cost = 223.236 km
From Sibiu          to Arad           -> Total cost = 446.473 km
Estimation          = 420.536 km
Priority Queue Cost = 867.009 km = (Total cost + Estimation)

From Arad           to Sibiu          -> Total cost = 223.236 km
From Sibiu          to Oradea         -> Total cost = 444.358 km
Estimation          = 434.745 km
Priority Queue Cost = 879.104 km = (Total cost + Estimation)

Possible paths:

From Arad           to Sibiu          -> Total cost = 223.236 km
From Sibiu          to Rimnicu Vilcea -> Total cost = 301.317 km
From Rimnicu Vilcea to Pitesti        -> Total cost = 348.536 km
From Pitesti        to Bucharest      -> Total cost = 456.108 km
Estimation          = 0 km
Priority Queue Cost = 456.108 km = (Total cost + Estimation)

From Arad           to Timisoara      -> Total cost = 48.459 km
Estimation          = 408.79 km
Priority Queue Cost = 457.249 km = (Total cost + Estimation)

From Arad           to Sibiu          -> Total cost = 223.236 km
From Sibiu          to Fagaras        -> Total cost = 287.59 km
Estimation          = 178.296 km
Priority Queue Cost = 465.886 km = (Total cost + Estimation)

From Arad           to Zerind         -> Total cost = 51.908 km
Estimation          = 431.034 km
Priority Queue Cost = 482.942 km = (Total cost + Estimation)

From Arad           to Sibiu          -> Total cost = 223.236 km
From Sibiu          to Rimnicu Vilcea -> Total cost = 301.317 km
From Rimnicu Vilcea to Pitesti        -> Total cost = 348.536 km
From Pitesti        to Rimnicu Vilcea -> Total cost = 395.755 km
Estimation          = 154.102 km
Priority Queue Cost = 549.858 km = (Total cost + Estimation)

From Arad           to Sibiu          -> Total cost = 223.236 km
From Sibiu          to Rimnicu Vilcea -> Total cost = 301.317 km
From Rimnicu Vilcea to Craiova        -> Total cost = 400.614 km
Estimation          = 183.042 km
Priority Queue Cost = 583.656 km = (Total cost + Estimation)

From Arad           to Sibiu          -> Total cost = 223.236 km
From Sibiu          to Rimnicu Vilcea -> Total cost = 301.317 km
From Rimnicu Vilcea to Sibiu          -> Total cost = 379.398 km
Estimation          = 213.803 km
Priority Queue Cost = 593.201 km = (Total cost + Estimation)

From Arad           to Sibiu          -> Total cost = 223.236 km
From Sibiu          to Rimnicu Vilcea -> Total cost = 301.317 km
From Rimnicu Vilcea to Pitesti        -> Total cost = 348.536 km
From Pitesti        to Craiova        -> Total cost = 452.104 km
Estimation          = 183.042 km
Priority Queue Cost = 635.146 km = (Total cost + Estimation)

From Arad           to Sibiu          -> Total cost = 223.236 km
From Sibiu          to Rimnicu Vilcea -> Total cost = 301.317 km
From Rimnicu Vilcea to Pitesti        -> Total cost = 348.536 km
From Pitesti        to Fagaras        -> Total cost = 458.356 km
Estimation          = 178.296 km
Priority Queue Cost = 636.653 km = (Total cost + Estimation)

From Arad           to Sibiu          -> Total cost = 223.236 km
From Sibiu          to Lugoj          -> Total cost = 397.029 km
Estimation          = 356.126 km
Priority Queue Cost = 753.155 km = (Total cost + Estimation)

From Arad           to Sibiu          -> Total cost = 223.236 km
From Sibiu          to Rimnicu Vilcea -> Total cost = 301.317 km
From Rimnicu Vilcea to Mehadia        -> Total cost = 461.891 km
Estimation          = 299.853 km
Priority Queue Cost = 761.744 km = (Total cost + Estimation)

From Arad           to Sibiu          -> Total cost = 223.236 km
From Sibiu          to Rimnicu Vilcea -> Total cost = 301.317 km
From Rimnicu Vilcea to Lugoj          -> Total cost = 504.328 km
Estimation          = 356.126 km
Priority Queue Cost = 860.454 km = (Total cost + Estimation)

From Arad           to Sibiu          -> Total cost = 223.236 km
From Sibiu          to Arad           -> Total cost = 446.473 km
Estimation          = 420.536 km
Priority Queue Cost = 867.009 km = (Total cost + Estimation)

From Arad           to Sibiu          -> Total cost = 223.236 km
From Sibiu          to Oradea         -> Total cost = 444.358 km
Estimation          = 434.745 km
Priority Queue Cost = 879.104 km = (Total cost + Estimation)

This is the shortest path based on the A* Search Algorithm:

From Arad           to Sibiu          -> Total cost = 223.236 km
From Sibiu          to Rimnicu Vilcea -> Total cost = 301.317 km
From Rimnicu Vilcea to Pitesti        -> Total cost = 348.536 km
From Pitesti        to Bucharest      -> Total cost = 456.108 km

Do you wanna try A* Search again? Yes or No?

A small change
One thing I changed in the code I posted on A* pathfinding search in C# - Part 2 was the foreach that enumerates the shortest path to write it on the screen. Before it read:

// Prints the shortest path.
foreach(Node n in shortestPath.Reverse())
{
    Console.WriteLine(n.Key);
}

Now it reads:

// Prints the shortest path.
foreach(Path<Node> path in shortestPath.Reverse())
{
    if(path.PreviousSteps != null)
    {
        Console.WriteLine(string.Format("From {0, -15}  to  {1, -15} -> Total cost = {2:#.###} {3}",
                          path.PreviousSteps.LastStep.Key, path.LastStep.Key, path.TotalCost, distanceType));
    }
}

As you can see I changed from Node to Path<Node>. To get this working I had to change the type returned by GetEnumerator in the class Path so that it returned Path<Node> instead of Node.

public IEnumerator<Path<Node>> GetEnumerator()
{
    for(Path<Node> p = this; p != null; p = p.PreviousSteps)
        yield return p;
}

This allowed me to enumerate over each path that composes the whole shortest path so that we can show the LastStep of the previous path and the LastStep of the current path. The Total cost travelled so far for each path is also available because we’re working with a path object.

Last note
A* is a really powerful search algorithm.

Hope you liked this series of posts about A* search as I liked to implement and write about it! It was a really good programming exercise.

Visual Studio 2013 Solution with C# Console Application Project
You can get the Microsoft Visual Studio Project at this GitHub repository:

https://github.com/leniel/AStar

To try out the code you can use the free Microsoft Visual C# Express Edition that you can get at: http://www.microsoft.com/express/vcsharp/