Project X Discussions

SOA Deployment Management

What's in a word.  In deployments the nature of the words you use can kill you.  SOA (Service Oriented Architecture) is living proof of this opportunity for confusion.

Remembering the fundamental guiding principles of SOA (see endnote below from wikipedia) we have an issue/opporunity around the concepts of interoperability, componentization, encapsulation and reuse.

To me this means that the service is independent in nature to the other services around it other than rules of interaction and to a lesser extend the platform.  So when you start talking about SOA deployment there is often confusion when reviewed or implemented by the school of thought of our old systems.

Since the systems are supposed to be independent entities that do not exist if not published in the registry, if I deploy a new service and don't tell anyone, do I need to test the whole system?  No!  If I add it to the registry do I need to test the whole system? No!

In the old world of tight coupling there were issues when you made even the simplest of changes which called for massive regression tests.

So when we think about deployment in a SOA world we need to break it up into different types:

1. Infrastructure
   • Upgrade/Add Hardware
   • Upgrade underlying enterprise platform that service lives on (BEA, SAP or Websphere)
   • Network
   • Location
2. Services Deployment - New
3. Services Deployment - Old - modifying an existing service, add elements or change business rules
4. New customers/consumers of our services

These to me are the main points of consideration that should drive behaviour, but choose your words wisely or you will spend so much time testing that you will spend more than your development budget and that defeats the purpose.

NetVault Backup and Restore Utility - First Impressions

I had a chance over the last two days to get some initial exposure to NetVault from BakBone Software.

So I was surprised, but in a good way.

This is a fairly straightforward product.  Some thoughts:

• Job Setup - Easy all click and select and then configure
• Scheduling - Very straightforward
• Plug-ins for various Enterprise systems - Oracle, Teradata, SQL Server, Postgres, file systems...
• Plug-ins allow for automated integration into the Enterprise system built in components.
• Operation - looks very easy
• Restore - only against already run backups.  Not able to set these up ahead.
• Monitoring - integration and ability to monitor from outside, logs and so forth looks limited - possibly use command line
• Integration with External job scheduling systems - possible command line available - could be cool

So far it looks like the marketing brochures promise.  And that is a pleasant surprise.

CoffeeCAST #44 - Floor Sweep Mentality and Details

Welcome to the 44th CoffeeCAST with Stephen Hayward of Project X Ltd.  Well it is Wednesday night before Graham goes off to sail and I grabbed him to chat about what is on his mind and have a quick CoffeeCAST.

Todays conversation was funnily I think more about the mind and attitudes or approaches to common issues in Information Technology.  Today though we only focused on the following:

1. Technology upgrades - The floor sweep mentality
2. Attention Detail
3. The life of Production Support

Download 44_coffeecast_44_floor_sweep_mentality_and_details.mp3

Listen in and we hope you enjoy and more importantly, hope you join in the conversation.  So drop a comment in the blog or send me an email and share your thoughts on the topic.

The audiocast is available on iTunes (as a Podcast) or here for download.  Have a great day and join the conversation.

Interfaces for Enterprise Systems - ODBC vs API

We have been working over the past several months with a client who has DataStage (IBM Information Manager).  At one point we reached a fork in the road and had a choice of which interface to use into the database - ODBC vs API.

Not sure exactly why we chose ODBC stage to connect to Teradata vs the API but we did.  We felt good about doing so.

So why am I wondering if this was still a good route.  It seemed to make so much sense, but we are having so much trouble keeping the DataStage .ini configuration stable to not have random Abend issues.  ODBC should be a perfect fit except when we use fast export or another more specific utility from Teradata.

OK so it should be fine.  The Oracle interfaces don't seem to have this issue.
It was and probably is fine.
But if it can not remain stable it is not fine.

Outside of theoretical rhetoric, has anyone else had that issue?

CoffeeCAST #43 - Collaboration in the Enterprise

Welcome to the 43r CoffeeCAST with Stephen Hayward of Project X Ltd.  I finally had a chance to catch up with Rex Lee who I have known for a while and have been bugging to come onto the CoffeeCAST and he finally submitted on behalf of the kids ;)

We took the time to talk about something which is top of mind to many corporations - collaboration in the enterprise.  Rex sets some framework on focusing on the client and then breaks it down into four components or success criteria:

1. innovation
2. speed / agility
3. holistic - end to end approach
4. employee engagement

In our discussion Rex mentions a couple of outside reference points:

1. McMaster World Conference on Innovation
2. IBM World Jam
3. Andrew McAfee

Download 43_coffeecast_43_collaboration_in_the_enterprise.mp3

Listen in and we hope you enjoy and more importantly, hope you join in the conversation.  So drop a comment in the blog or send me an email and share your thoughts on the topic.

The audiocast is available on iTunes (as a Podcast) or here for download.  Have a great day and join the conversation.

Communication, Understanding and Testing

Testing is often the ugly step sister of IT.  So it is often marginalized and sometimes forgotten.

Lately I have seen the opposite effect and that scares me as much as the other.  One of the indicators is in communicating what testing is required.  For me this is the dreaded word "Regression Test".

As systems evolve and the evolution of object oriented programming and service oriented architecture (SOA) evolve into the enterprise we need a new way to communicate the requirements for testing.  Let's take an example...

Systems Planning vs Monitoring

I have enjoyed some conversations as of late that have been showing me the difference between Systems Planning and Systems Monitoring.

These are two very different types of activities but very much intertwined.  Often the challenge is that when people start talking about planning about usage of an application they get caught in the IT metrics that come from monitoring and do not really give us a way to map these metrics or data points to the actual planning process.

This issue is really about the ability to link technology with the business.  The business view is the plan, which will think about how the user will grow in their use of the application.  The business does not care about the metrics really, but we can not answer the business question without a deep understanding of the underlying metrics and then be able to give a simple way to have someone be able explain what this will mean from a planning point of view.

Here are a couple of real examples, add some if you want:

• SOA - cost per service consumed
• Legacy Systems - MIP (millions of instructions per second)
• Business Process Outsourcing - Cost per transaction
  • Cost per minute in a call centre
  • Cost per call resolution
  • Cost per invoice process
• Databases - cost per gigabyte
• Backup - cost per gigabyte

In the above examples there are some very complex metrics that should include, people, process and technology cost both from a development and operational cost but if done properly really does a great job in helping quickly help the business in understanding in their terms.

CoffeeCAST #40 - Capacity Planning

Welcome to the 40th CoffeeCAST with Stephen Hayward of Project X Ltd.  Graham and I grabbed a quiet spot in the office with some fresh Starbucks (coffee for me and tea for him) to talk about Capacity planning.

Graham kicks off by discussing Mission statements which funnily enough gave us our segway into the topics of today.

• Capacity planning
  • planning versus monitoring
  • OLAP (online analytics processing)
  • OLTP (online transaction processing)
  • SOA (service oriented architecture)
• The model of on demand metrics
  • the old MIP model (no not this MIP)
  • IBM and on demand management

Download 40_coffeecast_40_capacity_planning.mp3

Listen in and we hope you enjoy and more importantly, hope you join in the conversation.  So drop a comment in the blog or send me an email and share your thoughts on the topic.

The audiocast is available on iTunes (as a Podcast) or here for download.  Have a great day and join the conversation.

CoffeeCAST #38 - Readiness and Master Data Management

Welcome to the 38th CoffeeCAST with Stephen Hayward of Project X Ltd.  We are on the road this week, having jumped to Ottawa and Montreal on a three city tour.  It is cold and sunny and having some fun.

Today I would like to Charles to the CoffeeCAST.  We are sitting here late at the lounge of the Sheraton in Montreal enjoying a non-Coffee and I decided to ask him to join in on a conversation that follows up some of the things we have been talking about.  Today we talk about:

• Readiness
  • FIDO - F*** It Drive On
  • The Dog and the Sharp Nail
• Master Data Management (MDM)
  Where we spend a bit of time on discussing it's place and so forth

Download coffeecast_38_readiness_and_master_data_management.mp3

Charles, thanks for taking some time today. 

Listen in and we hope you enjoy and more importantly, hope you join in the conversation.  So drop a comment in the blog or send me an email and share your thoughts on the topic.

The audiocast is available on iTunes (as a Podcast) or here for download.  Have a great day and join the conversation.

Collaboration in the Enterprise

One of the interesting things about social computing and it's evolution is the applicability in the Enterprise.  As organizations grow, the ability to communicate ideas and learning becomes more difficult.

The web has done wonders in enabling this sort of shared conversation and collaboration.  This morning we ran a quick netmeeting where we were interactively training some folks on application development of a new platform we built for ETL.  It was amazing to see - nothing new there right.  But how about if I recorded it and did an internal blog post on my internal blog - BOOM I have created free training material.  No cost, opt in.

The same is true about using blogs to help augment or replace forums in creating conversations and building bodies of knowledge.  Take for instance the creation of a new governance framework.  I can create a stream of group blogs to capture and create a conversation that can be leveraged internally to grow and share.

Now let's go and use it to create internal project lessons learned and the ability to interact with the information.  So I just finished a project.  I take 2 hours at the end to share on my corporate internal/external blog the high level thoughts around the project.  Now when someone searches they can offer up questions in the form of comments which someone can respond to, add information around.

All of these things are not new.  We have had forums and knowledge repositories.  But with the inclusion of RSS, google search and a number of other items that can be added in a mash-up search/knowledge portal we have a new way of collaborating in a cost effective manner.