How to optimally prioritize and schedule multiple migration projects?
i would like to ask, if anyone has any advice on what the best practice on the topic is?
The background: My company is looking to migrate software solutions of 50+ customers into a new data center. There is more or less about 3 Years time for that until the old data center is closing. Now how do i objectively determine in what order to migrate the customers? What is the best practice on scheduling?
I've done a lot of research but only find very vague concepts everywhere. What i found out so far is, that i can use a Scoring Model fe. Prioritization Matrix or Analytical Hierarchy Process to determine which projects to prioritize.
But that doesnt really fit my needs as im not trying to pick out the best projects but rather im looking to find a way how to optimally fit those 50+ projects into a timeframe of 3 years.
I hope someone can help me out here or point me in the right direction. Saving Changes...
You would first need to determine what you mean by "optimal". You can optimize things based on many characteristics like cost, flow time, disruption, resource utilization, weight, and any number of other things. It is very easy to sub-optimize based on one variable. That is why scoring models try to weight and combine multiple variables into a single number, so that there is some objective way to compare trade-offs.
If your question boils down to something like "we need to migrate the most important first", a scoring model would help identify what you mean by most important. If your question is more regarding efficient resource utilization, the answer would look very different. Without first defining what you mean by optimal however, all anyone can do is guess what is important to the answer. Saving Changes...
I would like to add the following for your consideration.
What is the level of risk / downtime during the migration? Is this the first time the migration has been done? Arrangement with clients is going to be important to schedule the timing.
Based on my experience, what goes first are be the ones with the least risk / impact of failure. At the worst case scenario, it is possible to revert back and avoid significant operation damage to the client. This may not always apply.
First is to understand the requirements & schedules for each of the migration project.
Resources and skill/competency level of the personnel is one of the key factors in deciding optimal schedule Saving Changes...
As it stands now, there are basically 3 types of migration projects.
1. Small standardized solutions that can be easily migrated (no costs, no project team) 2. Middle sized solutions with little legacy (low costs, shared project team) 3. Big solutions with huge amount of legacy (high costs, dedicated project team)
Right now there is a pilot migration going on, from which there will be templates and lessons learnt created for the following migrations.
Costs and resources are surely a factor, but secondary as of right now. The most important thing is to find a good schedule for the 3 years, so that we dont overload our capacities in the first year and also dont run out of time in the last one. Are there any tools, methods, practices etc. out there that could be useful for that goal? Saving Changes...
One method is called a sandpile chart and is very easily done in Excel.
Figure out what your resource profile looks like for each of the 3 solution types small, medium and large). Are they level loaded, bell curve, etc. and what is the duration? Your pilot program should help with that. Figure out what your resource capacity looks like the same way over the entire timeframe. Do you plan to ramp up your resources and taper off or use the same resource level start to finish?
Now you have a number of smaller curves that you need to fit into the area of your total resource curve, which is referred to as a stacked area chart or a sandpile. You move the various individual solution curves to different timeframes and see how to best align them for your desired resource capacity.
The planned resource curve is your "optimum", and by moving the projects around to fit the curve, you optimize their sequence based on the desired resource level.
In my experience, it works best when you start by distributing the largest ones over your entire timeframe, then fill in the big gaps with the medium ones, and finally the smaller ones although those can be easily adjusted over time based on your actual performance compared to plan. Saving Changes...