Information Collection
phase is based on initial interviews and discussions with the IT owners
of server consolidation & virtualization project aiming to:
- scope out the assessment expectations
- identify business requirements that will help in server
distribution on virtual machines
- consolidate the server list that will be configured for data
collection
- discuss system requirements to install “VirtualAssess DataCollector
Manager” and “VirtualAssess DataProcessing Manager” applications
- discuss data collection architecture if there are more than 200
servers to be monitor – details in product section
Data Collection phase
will give us all necessary data for the data analyze phase.
“VirtualAssess DataCollector Manager” is the application used to perform
and achieve data collection.
Collectors architecture design – for more than 200
servers we have to design a distributed collection. There is no upper
limit of servers to be monitor. “VirtualAssess DataProcessing Manager”
can aggregate and process collected data of 5000 servers with no
problem.
Collectors installation – install “VirtualAssess
DataCollector Manager” application
Collection parameters – domain account, admin account
for servers in other domains, number of days to run data collection
Collectors configuration – load server list, enter
parameters and run server connectivity check to ensure that all servers
are ready for collection
Start Data Collection – start collection and check logs
for unforeseen problems
Data Analyze phase will
process the collected data in order to be ready for server consolidation
scenario modeling phase. “VirtualAssess DataProcessing Manager” is the
application used to perform the following activities.
Data aggregation and processing activity will aggregate
server’s data calculating two values: hourly average value and hourly 95
percentile value. This will be computed for CPU, Memory, Disk IO, Disk
utilization, Network utilization parameters for each server. More
information in product section.
Generate top utilization reports for each of the
parameters: CPU, Memory, Disk IO, Disk utilization, Network
utilization.
Refine the final server list – depending on some
business requirements, some servers with high utilization of specific
application requirements may be extracted from the list of servers to be
consolidated.
Server Consolidation Scenario
Modeling phase is an cyclic phase that will give you the
capability of creating unlimited consolidation scenarios. “VirtualAssess
DataProcessing Manager” is used for this purpose.
Define Virtual Machines hardware characteristics
Define Scenario Parameters – depending on the
consolidation & virtualization technology, CPU overhead may be
introduced. Another parameter that should be introduced is server
variability which is a value that defines how smooth vs spiky (1 to 3
scale) is the CPU and Memory of a server
Run Server Distribution Simulation – simulate server
consolidation and distribution on the defined virtual machines.
Review Virtual Machines system load report after
migration – if the virtual machines load is not acceptable you
can manually and automatically redistribute servers in order to get the
desired server distribution. You can even add/delete/modify the virtual
machines hardware characteristics and re-run automatic servers
redistribution.
Deliverables phase will
contain the development of the final Server Consolidation &
Virtualization Report. This may differ on the business requirements
defined in the Information Collection phase. Deliverable samples will be
provided and described in the deliverable
section.
|
