Utvecklare

During the time as employed at StoneBeach there were two additional long term customers, where additional consultants from the company were brought in to help out when the work load was peaking, often due to releases or upgrades. At this point resources were brought in to support mainly in bug fixing and testing. Helen was a big help in these peaks, especially since she fast gets up to speed and she made difference for the progress together with her colleagues. In both cases the customers solutions were built on Oracle Web Center Content. 

Previous assignment was prolonged to continue the work to make EAB Toolbox more efficient and more accurate regarding data precision. 
The end goal is to make the batch job run quicker and automate it with Control-M to run on a daily basis, independently of the number of files. Today the batch job can handle all OpAMs of IKEA (around 7500) in one batch and it takes couple of hours, to be compared to previous limits around 50 files per week. 
Lots and lots of additional functionality is added, both new and changed features. This involves analysis and re-engineering of the code, mostly html, jsp and java, as well as of data structures (MS SQL). 
In the end it will give IKEA better control of existing environments and especially those that are not in use and should be decommissioned. This will be a cost saver as well as quality improvement. Other systems, for example LCM365, are using EAB Toolbox data as source of information and their data quality will as well be affected by changes in the EAB Toolbox. 
Inconsistent and wasteful data has been identified and to clean this out some adapted sql queries were developed and those should be executed by a DBA to clean up data whenever it is needed. 
Some more work remains to be done, like automate the creation of pdfs to correspond to the published visio files, add reports and information. Today pdf creation

To make sure IKEA could continue to use the system developed in previous assignment, the agreement was to prolong Helen’s assignment. 
NWP was going to order a big number of servers during their upgrade cycle and since the conclusion of previous POC was that the servers were delivered on time and with high quality, the decision was to continue to use this way of defining and ordering servers. This process change saved time and money for the big upgrade project and the outcome was highly appreciated. It also made it easier to keep track of the orders for the project management. 

As an add-on to this main task, the work to investigate the setup of EAB Toolbox was initiated. The root cause was that it came to IKEAs knowledge that there were data discrepancies in the system. When IKEA ran out of server names for Linux machines, they added an extra digit to the server name syntax. This resulted in unpredicted and unidentified side effects for surrounding systems, such as servers with 5 digits not being recognized. 
The goal was to find the root cause for the “5-digit server issue” and why servers with this syntax were not to be found in for example LCM365. Quite quickly the root cause was identified and fixed and the agreement was to dive into the code to find more bottlenecks. 

Another main issue with the EAB Toolbox was identified (not related to 5-digits issue), where the batch job was having performance issues. The batch job, where visio files are parsed and servers are extracted and mapped to placed orders, took too long time and was limited to run about 50 files per week due to performance issues. The consequences was that the data could never be considered fresh nor reliable.

In this assignment, Helen was brought in for her weblogic knowledge as well as her technical skills. Her social personality was fitting perfect in the role as single point of contact for a project named GLAS (Global Availability and Scalability) with the goal to identify the path for IKEA to fulfilling the need of availability and scalability. The project required one single resource to manage their tasks within OEDU (One Environment Deployment Unit) and supply them with the required servers for their tests. 
The need wasn’t as big as initially expected, so to fill up the time she was asked to investigate the order and build process to find the gaps in and low hanging fruits to improve delivery time as well as quality. 
The start off was to create verification scripts to enable the possibility to verify the deliveries of servers as soon as they were delivered by Build Service. This also decreased the turnaround time for faulty deliveries and in the end improved the both time and quality. To be able to execute these scripts, the input was manually created. The need to automate this creation led to starting the work to develop a way to define the build instructions in detail. 
Together with the project, NWP (New Web Project), an analysis program was started to define the infrastructure setup (the outcome was called Request template or later SBD (Solution Baseline Definition)). In parallel, Helen developed a complex “temporary” system was developed to be used as a POC and this work was done without funding’s (no money to buy or develop a permanent solution), so the tools used in development ended up being Excel and VBA. This tool contains the definition of NWPs (or any other solution’s) environment setup. Details like performance, software versions, monitoring packages, database details and so on are kept in the template. To define the order, the instance information is being added. The result after adding instance information is a complete definition of the infrastructure that is planned to order. When approved, the output will be for example different kind of JSON files (used by Ansible as input for building), verification input, summary for the different teams and so on. The purpose is to enable automation as well as provide one single source of information for the data consumers, independent of what shape of data they need. SBD, Helen and one IKEA colleague, also worked together with for example EDA and Ansible teams, to make sure we aligned with them as they were potential consumers of the 
The system was then used to order all upcoming environments and the outcome was close to perfect. Quality was better than ever and delivery time decreased. The need for rebuilding already delivered servers was reduced to nearly none and this saved time and effort for both MW-team as well as for the different App-teams. Finally, deliveries could be done with quality and consistency.

3 consultants from the same company staffed the small project, SDE (Software Document Engine), within the big program STMS (Supplying Transport Management System) at IKEA. One resource was assigned as Solution Architect and senior/advising developer and another one was handling infrastructure related tasks. 
As the third resource, Helen was the main developer. Since she at this point was a junior developer, it was of great use having a senior assisting developer as mentor to discuss and to act as advisory partner. She also learned a lot from acting as discussion partner for the solution architect. 
The system to be developed was going to create and handle the different transport documents used within global IKEA organisation. If these documents are not in place or are faulty, the fees are huge, so the cost savings for doing correct is enormous. 
The documents contained sensitive information, especially regarding costs and manufacture details, so the access model for the documents had to be extensive. This, as well as version handling, was handled by the document system and was adjusted to the different document types. 
The system was mainly developed on Oracle platform and used Web Center Content (for content management), Web Center Capture (for document scanning) and BI Publisher (for creating document templates). Tesseract was used to extract text from scanned documents and SOA was acting as a hub and orchestrator to connect different systems as well as being the engine for the whole solution. Since SOA is built on webservice calls, it was easy to enable SDE to receive calls from other solutions. 
The system was running on Linux machines with GIT and ClearCase being used for version handling of the code. 
Scrum and KanBan was the project method and the sprints were 4½ weeks. JDeveloper was used for SOA development together with a SOA plugin. PuTTY was used to access the Linux servers. 
When the project was getting closer to the end, the number of consultants was decreased. 
Helen was left as single consultant the last 8 months and acting as a Solution Architect as well as handed over to receiving SLF team.