As everyone including you may already know, switching from Probes/Sensors to Patterns is not advisable (basically).
I think this should be the very starting point for discussion.
Warning: DO NOT switch from probes to patterns if you are already running Discovery with probes, and your CMDB is already populated. If you do so, it is possible that the pattern Discovery process does not synchronize on the same values that the probe Discovery process does. This could result in duplicate CIs in your CMDB.
How do software models relate to Business Applications and Application Services? I am seeking to understand the data model and how to manage relationships for things like ServiceNow that we manage as a business application, but also need to track license usage. This brings up another question about how to manage SaaS with SAM.
When ServiceNow completes updating the conversion scripts, then there will be a path to convert from the Probe-based discovery to Pattern-based discovery.
If you already have an established CMDB using Probe-based discovery, moving to Pattern-based discovery will require planning. The classifiers will have to be updated and multiple ServiceNow created scripts will be run. The scripts that ServiceNow originally created were as follows:
Migrate Load Balancer to Pattern
Migrate Net Devices to Pattern
Migrate Windows to Pattern
Migrate Unix to Pattern
These were available in Kingston Patch 12 and London Patch 4, but have since been removed because they required additional updates. It is my understanding that the ITOM developers are working out the bugs in the scripts mentioned above.
I have been wondering this as well. According to the CSDM, Software Models roll up to Application Services, which themselves are Used by Business Applications.
The most consistent differentiator I have been told between Software and Applications is that Software is anything that can be discovered, whereas you make your own Applications and Business Applications.
Still alot there I would like to know about there about how they all work together.
A related question I don't think is in the material - how do we leverage, or DO we leverage software models in incidents? I could see that information being very useful for showing trends, etc.
I believe you can associate a Software Model with a Business Application and then attach that Application Business Service to the Incident. I'm checking in to see if you can associate a software model to an incident.
"Software Discovery Model" looks misleading name IMO.
"Discovered Software Discovery Model" or "Activated Software Discovery Model" seems to be better description of what it is. Because certain software discovery models appear only when appropriate software installations are found.
I'm wondering following questions
- Where is (Inactive) Software Discovery Model ?
- What if there are no longer applicable installations (because of uninstall, for instance), will Software Discovery Model disappear ?
Yes - patterns are newer. Patterns were integrated into the ServiceNow platform when ServiceNow purchased the Service Mapping technology from Nebula (Service Mapping originally was called Service Watch).
Originally, the rumor was that Probe-based discovery would completely disappear from the platform and completely replaced by Pattern-based discovery. Many customers had customized probe-based discovery and having to re-do all of their customizations into patterns would require a lot of work. The ITOM developers found a way for both Probe-based and Pattern-based discovery to co-exist with each other.
One of the key benefits of Patterns is that the Pattern Designer has a GUI front end that abstracts the developer from having to write JavaScript for customizations. Probe-based discovery customization requires writing JavaScript.
One other benefit of Patterns is that ServiceNow is providing more frequent updates to patterns:
It is never marked inactive. Discovery models can exist without an install associated. Even if tied via a discovery map to a software model without the associated installation record it will not impact your licensing position.
For the content library, it appears that there is not a way to get updates to the content library if you opt out of sending your data in... is this accurate?
If anyone wants to add the Normalization dashboard to their application navigator:
1. As a System Admin
2. Navigate to Software Asset menu
3. On the Overview module, click on the pencil
4. Right click on the header and choose Insert and Stay
5. Click on the Link Type tab and then in the Arguments, replace sysparm_dashboard id (in bold) to the ID of the Normalization dashboard (./$pa_dashboard.do?sysparm_dashboard=1d043526e7502300ba990558d2f6a915&sysparm_editable=false&sysparm_active_panel=false&sysparm_ignore_default_filter=false)
Attached is the schema I used to understand the ITSM asset management > software asset management structure. This schema is prior to the availability of the Foundation plugin.
We had some questions about the entitlement import and how it works to identify the potential duplicate entries.
Duplicate Entry Any entitlement imported that has the same PO Number, Publisher part number (and/or Software Model), License Metric, Purchased rights, and License Duration will be flagged as a potential duplicate. Simply verify that the entitlement is not a duplicate and click Create Entitlement on the Entitlement Import Error form.
If it is common practice to have multiple PO line items that purchase the exact same item, and there is potential for the quantity to be the same, a workaround would be to append a PO line item number to the PO Number field to avoid duplication.
How do people manage download locations and installation instructions? We were considering Knowledge, but we don't want the articles visible to the public without them going through a request process and being approved to download.
How do you allocate by things other than Users and Devices? Like per processor or per core, etc. I suppose that if they cannot be allocated, they can still be reconciled?
My company plans to implement SAM fourth quarter this year. Upon hearing the ideas in the class and how the system is put together I have made a rough draft of how we should go about this implementation. I would love any feedback.
IMHO, starting from "Step 1 - max discovery of all software" seems so different from most SAM practitioners do.
It might be better to determine the scope for certain publisher (1 or 2) & key titles for day-1. This is because not all the publishers are equal in terms of business impact.
For example, it is a very popular approach to prioritize the publisher which your organization spends the most . In most organizations, one/two among MSFT, Oracle, IBM, SAP will the become day-1 coverage.
So, if I were you I will make Step 3 comes at the very first. Step 1 and Step 2 are relatively technical, implementations related things. Planning is always important of course.
Without having the list of publisher/titles of what should be managed, we have no idea about which will be discovered without hassle, which will need custom probes/patterns to be inventoried.
But, there isn't 100% correct practice applicable to any organizations. This is from my little experience.
Optimizing software discovery is key. As well as defining the processes needed to ensure data quality. I have found it takes longer to gather entitlement information (Step 4) than one usually allocates time to during this phase of an implementation.
Typically, those are not allocated. The cores/Proc counts are utilized during the reconciliation activity. I'm happy to walk through this with you during a break.
We would want this to be done automatically as part of the fulfillment process. This would be sent via email automatically to the user with the license key that is provisioned.
For the CAL metric, reconciliation will use the max of "user/device allocations" or "total user/device CAL counts."
This allows organizations to still track individual user allocation for CALs, but also create CAL records based on direct/indirect access to a database.
Some mechanism must be created to extract the CAL consumption data, and this is highly dependent on which type it is, for example, it is super difficult to track CAL licenses for Windows Server earlier than 2012 R2; whereas it is typical to grant access to, say SQL Server, through an AD Security group. In the latter case, it is possible to write, say, a PowerShell script to calculate the number of devices and/or users (depending on the CAL type) and then an import-set and transform map to move that into the Client Access table (samp_sw_client_access).
With the Madrid release, you have the option of using the Citrix publisher pack to manage the information around the Citrix product. Also, you will also want to ensure your other publisher contracts allow for the distribution or use of software in a Citrix environment.
Since one of my prospect customer also uses Citrix, I ran into same question.
In case you use Linked-clone with Citrix, situation could be complicated.
As you know there are lots of deployment pattern in Citrix.
In case if you are to handle non-parmanent Citrix VDI, I recommend you to refer to this thread. In it, an independent ITOM consulatant guy gave me an brief idea about populate Citrix data into ServiceNow CMDB.
Please ignore me if you have already had a solution in handling your Citrix environment.
I was thinking about XenApp when I posed the question but you raise an issue I hadn't considered. Thanks for the link - I'd completely overlooked our VDI (permanent and non)
I agree, most would probably start at that planning phase. Unfortunately, the company I work for is a conglomerate of companies (5 major companies that continue to acquire others) and we don't have a concept as to what we even have in our environments. In our case, we will need discovery to even give us a hint as to which publishers we have a significant amount of installs.
As they have said in class several times when it comes to the answer as to how to do things, "it depends". I think you are correct for most companies that have a more mature IT department.
Do you use Upgrade rights to indicate if upgrades are included in the software agreement? Meaning that we do not need to pay to upgrade? We have some software where it is valid only for that version and we need to pay an upgrade fee to upgrade. Others there is no upgrade fee. If this is not managed here, then where is it managed? I would think we would need to differentiate between NEW purchases of the software at a higher version and upgrading existing licenses for a lower price. Also, how is that represented in the Service Catalog?
While creating the Software Model and selecting the Discovery Map
What if the discovery map doesn’t have results for a specific version
Example: SQL 2012, but selecting different Discovery Maps, none of them yield any results for 2012 (but discovery itself has found many sql 2012 installs)
Finding the root discovery map for SQL server works and can you can define the entitlement from that perspective
But it doesn’t allow you to manage the asset from a life cycle perspective as 2012 has an end of life support at a specific date
Question – are the Discovery Maps not being normalized properly? Or do you need to manually do something different?
