<< Real-world processes for a virtual age Forrester on IT automation to cut operational costs, increase service >>
Anxious about technical impact assessments? Application migrations giving you a migraine? Licence true ups getting you down? OK enough of the alliteration however the message is clear, if these IT operations problems are worrying you then you need to do something about it.
Before we go too far I will nip the CMDB in the bud right now because a CMDB for CMDB’s sake will only breed cynicism. Any configuration management solution needs to be just that - a solution, to a problem – otherwise it will add no value whatsoever.
So what is it that you can do to prevent all those downers, anxiety attacks and migraines? The single most important thing is to automate. Software license management and infrastructure and application dependencies are a technology problem so let’s use (the right) technology to solve the problems.
Asset and configuration management that relies on any manual process to discover assets and map dependencies will never be a reliable source of the truth. Virtualisation on it's own is equal parts solution and problem to maintaining a single source of the truth (I feel a headache coming on…). A data source that is unreliable makes people skeptical about configuration management, and we all know there are enough skeptics in the world!
Software licence true ups need not get you down if you are automatically discovering assets and software and have a service request process in place to handle installs and uninstalls. The most important things are: how many licences you have and how many are installed. Knowing that will make you feel better but to really pick you up, we suggest adding software usage will enable you to determine how many of those installs are actually being used so licenses can be reused on-demand.
Asset Management and Discovery Checklist
For automating software discovery, here's our minimum checklist:
- Automated discovery
- Automated auditing
- Can reconcile number of licences purchased and number installed
- Optionally provide for software usage / metering
Technical impact assessments for projects rely on correct data. This is why most are done manually as and when required, often going over the same ground many times over for different projects. Automated AND continuous infrastructure and application discovery is crucial to enable accurate impact assessments in a timely manner.
Infrastructure and application discovery is not much affected by complex applications and network topologies as the fundamental protocols used are well understood. The two main protocols are the Internet Control Message Protocol (ICMP) and the Simple Network Management Protocol (SNMP). These protocols have the advantage of being non-proprietary, supported by all leading manufactures and most importantly are low priority traffic that can be dropped with no impact to the network or discovery process.
Infrastructure considerations
Care and attention however is required with virtualized infrastructure as this is designed to increase the utilization of the physical infrastructure. Therefore any discovery MUST be unobtrusive and have as little impact on the network as possible. This is where agent-less discovery using the fundamental networking protocols are the best option, to avoid creating host and network processing overhead.
Single application to entire data center migrations are risky projects at the best of times and I have heard many a project manager say they are running blind on such projects. Application mapping provides a view into application communication and can visualize the infrastructure that applications depend on.
Providing accurate application mapping is crucial to removing the risk of migrations, including unforeseen capacity issues, security and authentication issues and unavailability of dependant systems.
Infrastructure and Application Discovery Checklist
- Automated discovery is agent-less
- Automated discovery uses ICMP and SNMP
- Can discover physical and virtual connectivity
- Can visualize the network
- Can provide or simulate impact assessments
One more thing. Once you have automated discovery in place then you can add future value by using the data for service management, particularly linking to the service catalogue and service level agreements. Typically this is done using the federation method to connect discovered data sources to a service management system.
Note however, federation itself can be risky especially if data exists that has been manually entered and/or updated. This can cause duplicates to be created and data being stored in the wrong location. Use the following checklist to ensure successful data federation:
- Can reconcile discovered items to prevent duplication of existing configuration items
- Can reconcile existing data to prevent data storage errors (provides easy data mapping between discovered and existing data)
- Can de-duplicate discovered items to prevent duplicate configuration items, this is especially important if two or more discovery tools discover the same item
- Can synchronise data, either automatically or regulated to identify unauthorised change via data discrepancies
- Can provide mapping and visualisation services
<< News < Real-world processes for a virtual age | Forrester on IT automation to cut operational costs, increase service >
post a comment
Post Comment