Andrew Meredith - Projects at Motorola

Why I left (Apr 2000)

This is an obvious question, to which there is a simple answer.

After over 8 years I decided it was time to move on.

IT Systems Design - (1995 - Apr 2000)

Due to my previous experience, from my arrival in the Software Development Group I had been acting as an internal consultant on IT related matters. In 1995 the Swindon ITS group allowed me root access to the Software Department Unix machines. This was a level of access previously reserved for ITS personnel alone and in fact up to the point that I left the company, I was still the only "non ITSer" that had root access to any Unix machine on the main network.

Around the beginning of 1998 the organisation decided to consolidate their GSM BSS Software development in Swindon. Among a great many other things this involved creating an IT system in Swindon to support this work. I was placed in charge of putting this together. By the time I left we had a system serving as all things for 75 developers that was considerably faster, more stable and cost less per head than its US equivalent. It was a Sun Enterprise E5500, with 8x400MHz processors, 8 GBytes of RAM and 200 GBytes of mirrored hard drive spread across 6 separate UW SCSI busses.

Due to my experience in IT systems design, I was called upon by a number of other departments within Motorola. As an example I was sent to the Motorola facility in Blackrock, Nr Cork in Ireland to set up a software development system for them.

As well as mainstream Solaris experience gained from the Motorola systems, I had also been building up experience of Redhat Linux in my IT lab at home. In late 1999, this started to bear fruit in terms of deployed systems at Motorola. A number of specialist problems were emerging in the communications labs. The PC equipment in the labs ran on Microsoft OSs, which were not capable of helping with these problems. I set up replacement configurations for these machines that played a major part in the quick resolution of the issues. Because of this, I was sent on the RHCE certification course to formalise my experience.

Member of ETSI SMG4 DGMH and 3GPP-CN3 (Nov 1997 - Mar 1999)

Given the experience I had built up on the CBC project, I was made an SMG4₁ delegate by Motorola and given particular responsibility for the CBC-BSC Interface, together with a colleague who's experience was of the BSC end of the link.

While attending the SMG4 DGMH meetings I was asked to become involved in the SMG4 GPRS meetings to help to clarify the interaction between GPRS networks and Internet Service Providers (ISPs). I came up with a new protocol called "Octet Stream Protocol" to enable simple interactions between GPRS users and Internet hosts. It was incorporated into the ETSI GPRS standard and is now a part of the GPRS system.

The responsibilities of the SMG4 GPRS committee were transferred to 3GPP-CN3₂. I was appointed by Motorola as a representative on this committee. Motorola then reshuffled its cellular infrastructure organisations and pulled all 3rd generation work to the US and I handed over to my stateside colleagues.

 

1

The European Telecommunications Standards Institute, Special Mobile Group is the body responsible for the creation and development of the GSM standard. It is the particular remit of SMG4 DGMH to oversee data messaging.

2

The 3rd Generation Partnership Project (3GPP) is a combined global standards group put together to come up with a single global cellular phone standard called UMTS. Core Networks Subgroup 3 (CN3) is the committee responsible for data services.

Java based Cell Broadcast Centre Prototype  - (Apr 1996 - Feb 1998)

It was felt that the Solaris/C++ based CBC Simulator was too expensive to deploy, given the high cost of the hardware. It was therefore decided to rethink the platform and the architecture. It was rewritten, with an improved structure, in Java on a Windows NT platform. At least it was mostly Java, there was a C++ device driver for X.25 connectivity using JNI. The use of Microsoft's own Java was avoided for compatibility reasons. This was once again a 2 head project, but this time it was even further in the background, due to both team members being regularly diverted onto other shorter term tasks.

In Feb 1998 the Java CBC was deployed on the GSM World Congress Motorola MCellAccess network. This was a temporary network that was connected to the France Telecom Mobile live system for the duration of the Congress. This made it the first live CBC in Europe, the home of GSM. It also sent the first CBC controlled Cell Broadcast message in Europe.

The package is also used to test and debug the SMS-CB feature on the Motorola BSC, so it has to be updated on a regular basis as the CBC-BSC interface standard evolves.

Cell Broadcast Centre Simulator - Prototype Phase (Aug 1995 - Mar 1996)

My manager also manages the Motorola Base Site Software development operation in the UK. His BSS team were developing the Short Message Service Cell Broadcast feature. This feature requires an external controller in order to function. At the time the project started there were no CBCs in existance. It was decided that myself and my CDFE colleague should develop a Cell Broadcast Simulator for use as a test tool for the Base Site feature. The first working prototype was based on Solaris 2.4 with the Sunsoft SunLink X25 protocol stack. It was written in C++ (95%) & C using the Sunsoft C++ Visual Workshop. It had a client program with a Motif GUI that was partially written using IST's X-Designer.

Customer Data Front End - CDFE (Jan 1995 - Jul 1995)

One of our major customers required a way of taking the system configuration data that they already had in their own internal format and automatically reformating it for use with the DataGen product. It was decided that myself and the other remaining DataGen team member were the most appropriate people to take on this project. We were both placed in purely technical roles with our manager taking on the UK based project management responsibilities. Once a formal definition of the features and architecture had been completed enabling a lifecycle and therefore cost to be estimated, the customer decided that they did not require it.

DataGen - As a product (Jun 1993 - Dec 1994)

The productisation phase involved porting to SunOS 4.1, a deal of formal documentation and a larger amount of interdepartmental coordination. The number of staff grew to a peak of 6. I was the "Technical Authority", there were 3 contractors and 2 full timers. Once the team had completed what was considered a full feature set, it was decided to move the project to Cork in southern Ireland in order to integrate it with the "Operations and Maintenance Centre". DataGen passed through it's last UK set of Systems Integration tests and First Office Applications tests with a clean bill of health.

DataGen - As an internal tool (May 1992 - May 1993)

This project started as an internal tool for automating the task of creating systems config files. The platform was a Motorola 88000 based SVR3.2 UNIX mini with the product using Motif and Informix Online. Myself and one contractor shared the task. It was originally to be used by two of the systems engineers. As time went on the project grew and at the end of this phase there was myself and 3 contractors on the development and we had been given 2 overseas customers. It was then decided to productise DataGen and make it a formal Motorola product..

PC Revgen (Feb 1992 - Apr 1992)

This project was to create a reverse engineering tool that would take a 1.5 MByte, 68000 based C "struct" file dump, uploaded from a Motorola Basesite, place it in memory in 64 KByte (or less) chunks, while relocating pointers and from this recreate the original configuration information. The platform was a portable PC running MSDOS, hence the need for 64 KByte segmentation.

Side Projects

WWW cross functional team - InfoNow (Mar 1995 - Aug 1996)

I had been experimenting with the NCSA HTTP server on my workstation and so was invited to join a small cross functional team that were trying to kick start a division wide WWW service. My role on the team was technical, I was one of 3 who could produce active web content and help train people in the creation of web pages. Motorola Swindon now has an encryption enabled Netscape Commerce server running on a Sun SPARCServer 1000. It now boasts an average of 15,000 hits a day.

Web Based Phone List (Jun 1995 - Jun 1997)

At the start of InfoNow team it was determined that there needed to be a "Killer App" to get people using the web. I took the prototype Informix 4GL based system and produced a production quality system based on C++ and Informix ESQL/C. It allows the user to update any entry, placing their edits in an intermediate table for review.

It was eventually taken offline once management were convinced of the usefulness of such a system and brought a new global system online.

Web Based Customer Intelligence System (Short project complete Aug 1996)

I was commissioned (on cross charge) by the marketing department to produce a web based customer intelligence system. It has password based access control, to allow different groups of people, access to different levels and flavours of information. It is entirely based on bourne shell script, running under Solaris 2.4.

This page since Jan 2000  Last Modified Thursday, 01-Jun-2006 11:42:34 BST