In the previous two parts of this series I have discussed the following
In the third part in this series I intend to discuss Solution Design, now this isn’t going to be an in-depth technical discussion on how your solution should be designed or configured. More what to expect during the design process some elements you may wish to look out for and questions to ask.
So by know you will have a clear understanding of what you want to achieve, your shortlist of partners will also have understood these requirements, you may have discussed some ideas of the technologies you want to be looking at and you may have seen some of these technologies demonstrated to you.
A hugely important part of the design process is data collection. Without this the whole solution design will be finger in the air. Whilst the people you are dealing with are likely to have a lot of experience in the area, nothing will make up for doing proper performance analysis. This will ensure that a solution designed for you has been designed to your needs. There are numerous different ways this data will be collected an analysed, there maybe a cost for this process due to the amount of time this process can take.
VMware partners will be able to offer you VMware Capacity Planning, this can be presented in two formats a Consolidation Estimate and the more in-depth analysis and report a Capacity Assessment. For more information about these please visit VMware’s Capacity Planner webpage. >> http://www.vmware.com/products/capacity-planner/overview.html
Platespin Recon is also another tool that maybe used for capacity planning for more information visit their website here >> http://www.novell.com/products/recon/
Finally sometimes nothing beats good old perfmon statistics, during the design stages you maybe asked to collect some key statistics from your servers, I particularly use this method for grabbing IOPS statics from physical servers that are due to be virtualised. This means that I am able to adequately size a customers SAN requirements for their current infrastructure and future growth.
There are of course many more methods and products on the market that maybe used to gather these statistics.
What is most important during the statistics collection is that for the period the data collection is ran, represents a the full range of use for your system. For example if your data collection is only ran for the first two weeks of a month and at the end of the month you run reports and payroll, these key areas of performance utilisation will have been missed from the analysis. For this reason data collection will often run for 4 – 5 weeks.
After the performance data has been analysed you will be presented with a report and a final design. In my opinion it is important that this is done face to face so you have an opportunity to discuss the design with the solution architect and raise any questions about the recommendations.
Below are a few areas you may wish to check or discuss
- Levels of VMware licensing, understand the different levels of licensing and what you are and aren’t getting as part of you suggest solution >> http://www.vmware.com/products/vsphere/buy/editions_comparison.html
- Redundancy – A big reason for virtualising an environment is to improve the redundancy for the systems hosted on the environment. If this is an important aspect for you (And it usually is in a production environment) you want to ensure that all the elements contain high levels of redundancy. These include multiple hosts (with multiple PSU’s, HBA’s NICs etc) the use of VMware HA, multiple switches for path redundancy, multiple controllers in the SAN. In some environments if up time needs to be guaranteed this may also include multiple SANs with synchronous replication. I have seen too many times solutions by competitors trying to make a quote seems cheaper than the competition by leaving out these levels of redundancy particularly single controller SANs!
- How does the recommended solution meet the needs of the performance and capacity statistics
- Why have the different products / technologies been chosen and what do they offer you that other products / technologies don’t
- Has future expansion been taken into consideration and what is the upgrade path when the time comes to add more capacity
- What training is going to be involved for your IT administrators to be able to manage the solution in the future.
- What support levels have been included with the various products.
- If DR is part of the solution, how will it work, what is the recovery period and what is the scope for potential data loss in a DR solution.
- If DR isn’t included at this stage how could it work in the future?
- If backup is included how will it work, how many recovery points will you be able to keep online, is there means of archiving included if needed.
- Has the bigger picture been taken into consideration, what effect does this solution have with other systems, processes or elements within the company
- If monitoring is required has it been included and what levels of monitoring does the chosen product give you.
- What are the limitations of the recommened solution, are these an issue?
Off course there are a infinite amout of details you may wish to cover, the above list is only meant to give you a starting point.
Make sure that you are comparing designs from different partners on a like for like basis, ensure they have understood your requirements and that the designs reflect this. If there is a huge cost difference between partners find out why this is, what does the more expensive solution offer you and do you need those added features?
At this point if you haven’t already seen demonstrations ask to see demonstrations of the various products. Ensure that you are completely comfortable with what has been recommended, if not tell the partners what is missing and ask for those elements to be included in the design.
Again I am going to leave this post here and will add to it over time, if you have anything you wish to contribute please let me know barry (@) virtualisedreality.com