Gartner’s new report, Software Vendor Auditing Trends: What to Watch for and How to Respond (May 23, 2012) contains  interesting food for thought for enterprises.  According to the report,  65% of the respondents in a recent Gartner survey said they had been audited by at least one software vendor during the past 12 months.  The vendors reported as auditing for software license compliance most frequently were (in alphabetical order) Adobe, Attachmate, Autodesk, IBM, Infor, Informatica, Microsoft, Oracle, SAP, Symantec and VMware. 

The report goes on to point out the myriad license compliance issues of each of the high-audit vendors, along with Gartner recommendations for how to best address these issues.  A close reading of the Gartner “primary area of risk” portion of the report highlights a common thread that enterprises must take into account in order to be “software audit ready”.  Here are a few short snippets:

• Adobe: “…Inventory discovery tools cannot ensure compliance with all license terms and conditions, such as the contractual restriction that Adobe Acrobat forms may save no more than 500 completed responses without a server product license, as these tools identify only the software installations and not the way in which the software is used...”
• Attachmate: “…In many cases, customers are confused about the difference between the mainframe and thick-client (standalone) versions of the products, and about the ways in which they can be used…”
• Autodesk: “…Autodesk's license models are complex and confusing, particularly the way in which subscription entitlements and prior version rights work… Autodesk products are notoriously difficult to audit and many clients find that their incumbent tools are unable provide the necessary information to demonstrate compliance with negotiated license terms and conditions…”
• IBM: “…IBM has announced and set forth a new International Passport Advantage Agreement (IPAA), effective 18 July 2011, with changes to terms and conditions… When reviewing [license]entitlements, ensure that these terms are updated in your contract and SAM databases and processes to track and manage compliance…”
• Infor: “…Risks include failure to comply with the usage rights for software licenses, source code and object code, for software licensed now or previously by Infor, Baan, SSA Global, Lawson Software and other licensors' products acquired by Infor… Infor customers are often not aware of these limitations or issues as IT asset management and IT procurement may not be actively reviewing these older license agreements…”
• Informatica: “…Customer [software] audits seem consistently to target several key areas of Informatica functionality. Because PowerCenter (and many Informatica products) are licensed on a per-CPU-core basis, customers need to ensure they are not deploying the software on more CPU cores than they are licensed for…”
• Microsoft: “…The most common compliance issues faced by Microsoft customers typically relate to a lack of understanding of the subtly nuanced licensing requirements…”
• Oracle: “…The primary areas of risk appear to be vague areas of license metrics and usage entitlements, including changes in license metric definitions that may show up in ordering documents that differ from those of some earlier deals…”
• SAP: “…SAP license bundles have shifted considerably over the years. The exact bundle was generally not well defined in contracts, which now makes it difficult for clients to know whether they are entitled to use specific software in their agreement…”
• Symantec: “…Both desktop and server virtualization require a detailed check of license terms…”
• VMWare: “…An ELA (Enterprise License Agreement) grants licenses for software to be deployed and used before the ELA expires. Customers who set aside licenses just in case they might be needed in future can find that these licenses are not covered under the terms of their current ELA, especially if it is uncapped…”

Even a quick reading of this Gartner report reveals that enterprises cannot hope to be compliant and software audit ready if they only concentrate on inventory and counting of licenses.  Indeed, the greater risk comes from not being aware of the complex licensing terms and product use rights contained in each software license agreement, and ensuring that users are at all times compliant with those terms.

Gartner provides excellent recommendations on the actions that enterprises should take to ensure compliance with those license terms.  The question remains whether those recommendations can be put into action without automation.  Given limited time, resources and the enormous complexity around software asset management and license optimization, I think not.  Having automated license management solutions in place capable of tracking license terms, applying software product use rights, and reconciling software installations and usage with license entitlements would seem to be the only practicable solution to assist the resource-challenged software asset manager implement Gartner’s recommended approach.

***

Readers may also be interested in this recent blog post on the same topic. To learn more about Managing IBM License Complexity, please download our white paper.

As the cloud becomes the more prevalent and preferred architecture – having a succinct, automated process between Development (Dev) and Operations (Ops)— DevOps—will be a critical success factor for IT shops.  During my research for Visible Ops Private Cloud and Client4Cloud – it became clear that IT shops that automated bad processes or did not have processes in place – only got into trouble faster.  The less process and more rigid controls IT put in place for the business the more the business would work around IT to circumvent policies that inhibited productivity.

The agility, flexibility, and dynamic nature of cloud computing provides significant ROI to the business but only if IT can create and automate processes that provide compelling value to overall productivity of the company.  As IT Ops becomes the "Cloud Broker" and is able to create virtual environments within minutes – Dev will need to automate the packaging, handoff and analysis to keep up.  One individual interviewed for Visible Ops Private Cloud indicated that although it was great that IT could  provide the virtual environment within 15 minutes it still took them over 9 days to have critical applications installed and running due to poor processes for software licensing, procurement, analytics, and change management.  (2011, ITPI)

Many senior development and operations executives I have spoken to are focused on – how to create a cloud environment that provides Infrastructure as a Service (IaaS) for the business. It became apparent during many discussions that DevOps is the next logical step.  Meaning, once the infrastructure is in place and automated – business will require IT to automate the rest of the process to reduce the overall time to value for the organization.  Successfully automating the handshake between Dev and Ops will be a critical success factor in companies realizing the true value of cloud computing.

Key elements to consider on your DevOps journey:

• Development workflow that enables collaboration across multiple geographies, groups, and departments.
• Secure source code repository with branching that also have a sound disaster recovery and security controls to reduce mistakes and malicious activities.
• Automated packaging and provisioning of applications across physical and virtual paradigms leveraging current backend systems for Asset Management, Configuration Management, Change Management, and Service Desk.
• Automated test and dev that are cloud enabled to provide both depth of scale but also ability to burst across various cloud environments.
• In depth analytics on application and feature level usage along with software license entitlements to ensure that not only technically can you deploy in the cloud but legally can you.
• Automated software licensing handshake across hybrid clouds – also known as licensing as a service to enable lifting and shifting of workloads without the drag of manual approvals for bring your own software licensing policies of 3^rd party cloud providers and ISVs.

DevOps is coming of age and will be more critical for both development and packaging moving forward.  The key is automating the handshake and processes now as part of the overall cloud implementation.

Want to learn more about DevOps…check out these upcoming events where Paul Peissner from CollabNet will share DevOps insights and how Flexera Software solutions fit within the process.

By John Emmitt

Everyone who is deeply involved in software asset management (SAM) will tell you that (a) its not a quick fix type of thing—its got to be an ongoing program, (b) its mostly about having good, indeed best practice, processes and skilled people (but having robust, proven technology can make a huge difference in terms of successfully implementing those processes—more on this later), and (c) it can be hard to build a business case for implementing a SAM program and get the support of senior management. Why is that?

The Gartner list of top 10 CIO tech priorities for 2012 shows that SAM and License Optimization are not (specifically) on the list:

1. Analytics and business intelligence. (Last year's rank: 5)
2. Mobile technologies. (Last year: 3)
3. Cloud computing, including SaaS. (Last year: 1)
4. Collaboration/workflow technologies. (Last year: 8)
5. Legacy modernization.
6. IT management. (Last year: 4)
7. CRM
8. ERP applications
9. Security
10. Virtualization. (Last year: 2)

If you look back at previous years, I’m quite certain that software asset management and license optimization weren’t on the list then either. We do see “IT management” on the list, and maybe this is where we need to assume that SAM and license optimization are really “showing up”—SAM programs help organizations manage IT assets— to gain control and improve visibility of those assets and have better financial management of the assets.

The same Gartner survey that produced the above list also looked at top business priorities, and in third place they have “reducing costs”. So, we see that cost reduction is still a high priority. And guess what? Software asset management and license optimization programs can help organizations reduce costs for software, not to mention helping to avoid unbudgeted software audit true-up expenses. Furthermore, software typically represents 20 to 35% of the IT budget—it’s a big chunk of the CIO’s budget. So, again, why isn’t it on the list?

I have a couple of possible explanations for this. First, traditionally, software asset management technologies and projects focused on counting assets—hardware and software inventory and maybe doing some reconciliation against purchases. But the process was not very automated and was far from easy—it was a lot of work to bring in the license entitlement (purchase order) data, figure out what was actually purchased and match that up with the inventory data. You may have purchased Adobe Creative Suite 5 (CS5) Design Premium, but your inventory data showed that you had Acrobat and Dreamweaver on some machines and Photoshop and Illustrator on some others, for instance. It took some real digging to figure out that the Creative Suite purchase covered some or all of those installations. At the end of the day, maybe all this hard work got you through an audit or annual true-up event, but it didn’t help you dramatically reduce software costs or avoid the audit in the first place.

Second, customers who have been successful with SAM programs are almost always reluctant to tell their stories. So there is not a lot of verifiable data out there that says, if I invest X in SAM and license optimization, I’m going to have an ROI of Y. The reasons for this include the fact that these organizations have to continue to do business with the software vendors, even after they have achieved significant cost savings with that vendor, and they don’t want to rankle the vendor by publicizing the savings. That’s understandable, but also unfortunate from the standpoint of advancing the cause of SAM in the marketplace. Its also ironic, because the enterprises have been over-spending on software with these vendors for many years in some cases.

To address the first issue, I can say that SAM and license optimization technology has come a long way over the past few years. Tools now have built-in Stock keeping Unit (SKU) libraries, for example, that allow purchase order data to be automatically matched up with inventory data to help streamline the reconciliation process. Now you can see your license position at “the push of a button” and maintain continuous license compliance—avoiding those license audits. License optimization tools provide product use rights libraries that capture vendor license entitlements, which, when properly applied can reduce your license consumption considerably—right of second use, multiple (versions) installation rights, virtual use rights, cold backup/failover rights, etc. all entitle you to have multiple installations covered under a single license. This type of optimization was hard to do before, and as a result, companies over bought software licenses.

With regard to the second issue, we have many major enterprise customers that have successfully implemented SAM and license optimization programs and have reaped the rewards in terms of—increased control and visibility of their assets, reduced license liability risk, and significant cost savings. The actual savings is typically between 10% and 30% of their software spend in the first year of implementation. One Global 500 customer with more than 100,000 employees, for example, has realized a savings of more than $35 million on just two major vendors. That equates to about 17.5% of their total annual software spend of $200 million. In several cases customers have achieved reductions of 50% to 60% of their ongoing software maintenance costs for major vendors such as Adobe and Oracle.

Building the Business Case

Let’s build a rudimentary business case for a software asset management and license optimization program for a hypothetical financial services company that has $5 billion in revenue. The typical IT budget as a percentage of revenue varies by industry, with financial services coming in at about 4% of revenue:

For our hypothetical Financial Services company

Total IT budget:  $5B * 4% = $200 million

If their software budget is say, 25% of their IT budget, then:

Total Software budget:  $200M * 25% = $50 million

If the company can realize just a 10% savings on its software spend by implementing a SAM and license optimization program, the savings is:

@10%: Savings from SAM and License Optimization: $50M * 10% = $5 million

Let’s estimate that the first year cost of their SAM program, including IT staffing, consulting services and tools is about $850,000. Then the return on investment in the first year of implementation is:

ROI @10% savings rate: $5,000,000 - $850,000 / $850,000 = 488%

We see that the return on investment is substantial, even at the 10% savings rate. And we have empirical evidence from existing customers that this savings rate is very realistic and in fact is at the low end of the range. Now what CIO doesn’t want to invest in a program that returns more than 400% in cost savings and reduces license liability risk at the same time?

***

To learn more about getting started on a Software Asset Management program, view our on-demand Webinar.

Flexera Software, the leading provider of Application Usage Management solutions for application producers and enterprises, announced today that its market-leading enterprise license optimization solution, FlexNet Manager Suite for Enterprises, won the Software & Information Industry Association (SIIA) CODiE Award for best asset management solution.  This is the third CODiE award win for Flexera Software.

SIIA, the principal trade association for the software and digital content industries, announced 28 CODiE Award winners in different categories on May 10 at a ceremony in San Francisco.  A CODiE Award win is an especially prestigious honor, as each award winner was first reviewed by a field of software industry executives, whose evaluations determined 119 finalists. SIIA members then reviewed these finalists and voted to select 28 CODiE Award winners in different categories, including FlexNet Manager Suite for Enterprise’s win as the best asset management solution.

FlexNet Manager Suite for Enterprises has emerged as the leading solution allowing large, global enterprises as well as midsize organizations to move up the maturity model, going beyond software asset management (SAM) to achieve continual enterprise license optimization. The solution automates entitlement based license management and reconciles software purchase order and contract information with actual application installation and usage data, to reduce license, maintenance and audit costs while maintaining license compliance.  As a result, enterprises gain unprecedented control of their software estate, driving as much as 5-30% savings in annual software spend.

Read the full press release here.

When an enterprise decides to take on in earnest the challenge of software asset management (SAM), it usually begins with a project.  The enterprise usually has some pre-existing rudimentary tracking in place, even if only entry of service requests for software installation into a spreadsheet.  The SAM project team will examine the requirements and capabilities for software inventory and entitlement tracking in the organization, select a product and/or service to meet its objectives, and implement the new solution in the enterprise environment.  However, one critical element is often overlooked:  conversion of the SAM project, after completion, to a sustained SAM and license optimization program. 

It’s understandable that enterprises are often focused on the SAM project.  The project is what should drive process implementation and improvement in the enterprise.  The SAM project is aligned with the enterprise’s business goals; as part of the project, an enterprise must understand and design the right business processes to achieve these goals.  The project phase is also when infrastructure to support SAM is deployed, when software inventory collection systems are integrated, and often when initial software reconciliation results are delivered.  In some cases, the project might also entail some level of license optimization in which the organization’s license entitlements, including product use rights, are applied to minimize license consumption. This could be done for certain key vendors in the initial project phase, for example. This optimization process goes beyond traditional SAM and is a key requirement for maximizing the ROI of the project and/or the ongoing program. A SAM project is full of activity, usually has high visibility, and because it represents change, is closely monitored. 

All good things must come to an end, and SAM projects are no different.  By their nature, SAM projects are short-lived.  This means that for the enterprise to achieve lasting benefit and a mature SAM environment going forward, the foremost goal of the SAM project must be the establishment of a SAM and license optimization program.  Such a program delivers ongoing value to an enterprise simply by continuing and improving on the processes and technologies established during the SAM project.  It sounds straightforward, but often this step is overlooked by enterprises.  Of course, in many cases, some or all members of the SAM project team will be members of the SAM and license optimization program team.  However, in some enterprises, particularly enterprises with significant IT outsourcing, at the end of a project, many resources will be redeployed onto other projects.  Without a transition plan, the newly minted SAM environment will be neglected and will grow stale over time.  Up to date reports will not be available, and ultimately the enterprise may need to create another SAM project to refresh the environment. 

Therefore, in order to create sustainable SAM gains for the enterprise, a SAM project must consider how the ongoing program will operate.  In particular, the project must forecast staffing requirements for ongoing license management activities, including maintenance of SAM infrastructure, software upgrades and issue resolution, and especially ongoing license reconciliation, compliance reporting and license optimization.  Do not assume that the time required for these activities will not change from current processes.  If an individual is currently tracking license allocations by manual spreadsheet input, it should take less time for the same person to do that using a more automated system, but they may have other more advanced reports to generate and analyze going forward. This could include license compliance reports, trend analysis reports and predictive analysis simulations to measure the impact of IT environment changes on their license position, to name only a few.  The value of this more-mature SAM and license optimization analysis is certainly worth it to the enterprise, but it does not come for free and must be staffed to be realized. 

In order for any enterprise to realize long-term benefits from a software asset management project, that project must be established as an ongoing program with appropriate staffing and support.  An established SAM and license optimization program will allow the enterprise to realize a higher return on investment on SAM processes and technology, not just for three months after the project conclusion, but year after year. 

***

To learn more about best practice processes for establishing a SAM and license optimization program, please download our whitepaper: NCC Guideline for IT Management: Software Asset and License Management Best Practices.

Are Random Processes Really Random?

Today’s blog is really a thought exercise on random processes. Do random selections always yield random results? Let’s run a test.

Let’s say that we have a Cloud compute grid where the servers are mapped to a 0-1000 by 0-1000 grid. Furthermore, we are using XWare virtualization technology and  xMotion to move workloads around the compute grid randomly. Each time a workload moves, the direction is randomly selected from (N, S, E, W, NW, NE, SW, or SE) and the color assigned to each. When the workload lands on the next server, the server is color coded. The distance traveled is always 2/3 from where it is to the edge of the grid, where the edge of the grid is determined by the random direction.

In the example below, the workload starts at randomly selected X and moves South (direction selected at random), lands on a server, which is colored purple. The next direction selected at random is Northeast and the server where the workload lands is colored blue. The next random move is Southeast and the server is colored green.

This compute grid is really big—1001 by 1001 computers or 1,002,001 computers. To see if the resulting workload moves result in a random pattern or not, this process was repeated 1 billion times, with the picture (a fractal) below being generated. We see that a seemly random process can lead to a very deterministic result. Interestingly, only 62% of the servers were touched by this workload.

In an IBM audit situation, if you could convince IBM that you should only pay for 62% of the servers in the cluster, because you can prove that xMotion will only move the workload to this percentage of servers, you would still lose the argument, since XWare isn’t an IBM Eligible Virtualization Technology!

Even if it were an approved virtualization technology, you’re probably still on the hook to license for all the servers in the cluster if you have “full capacity” licenses (depending on the license model in use). If you have “sub-capacity” licenses then you’d have to license for the Virtualization Capacity of the group of servers. Virtualization capacity is “the sum of the virtual core capacity available to a product.”

OK, so I managed to get a little bit in there on license management.

***

To learn more about how to simulate changes to your IT environment to understand the impact on your license position, please visit our website to read about ‘What If’ analysis.

One problematic area for Citrix XenApp (and other application virtualization technology) customers is how to ensure they buy the correct numbers of software licenses for products they deploy and use from their Citrix servers.  While the licenses they buy from Citrix are quite straightforward, the licenses for virtualized applications that they use in this environment can be troublesome to count, particularly if some users are also using the apps locally on their desktops as well.  Since every third party vendor would look on this kind of deployment differently, enterprises need accurate methods to count each type of license. License optimization tools that help count the usage and license consumption can be very beneficial and help avoid over-counting.

There is often an overlap between licenses used in a Citrix environment and those used in other end user environments, such as the traditional desktop. For instance while a Citrix farm might make MS Office available for all users, some users may already have a local install and consume a license for that device, which also covers them for the Citrix virtualized application delivered to the same device (desktop, laptop, etc.).  The examples below illustrate that depending on the license model and usage scenario, different numbers of licenses are consumed. In the first case, a User based license will consume only 1 license for 2 devices, whereas a Device based license requires 2—one for each device. In the second scenario, only one license is required to access both the local installation and the virtual application, regardless of whether the license is User based or Device based.

Figure 1 - Examples of Licensing Scenarios in a Citrix XenApp Environment

Why is this important?

A customer with just under 5,000 employees, including about 1,000 users who also access their software via Citrix, was recently trying to calculate their license consumption under all of the above scenarios. For some titles they were getting numbers greater than 5,000, even though users were meant to be able to leverage their license for a locally installed copy.  The numbers were obviously wrong, but how could they work out an appropriate license level? In addition, some of these titles weren’t used by all users, so just buying one copy for each employee would have been expensive and also the wrong approach.

They needed to come up with a better way of calculating the consumption of licenses for a combination of locally installed software and virtualized applications, without wasting money on unnecessary additional licenses.

How is it different from a traditional software environment?

In a traditional software environment, companies are used to the problem of just counting the number of software installs on the physical hardware. In a Citrix XenApp virtualized application environment, the application software isn’t installed on the end-point device, so it doesn’t show up in inventory. To solve this problem, some organizations are using next generation software asset management and license optimization products to collect application virtualization data from the Citrix server—for example, information on users with access rights to the virtualized applications and possibly, the actual usage data, if the Citrix Edgesight tool is also in use. This information can be combined with regular inventory from the end users’ machines to find out:

• Who has a local copy of the application on their desktop or laptop
• Who uses the software via Citrix XenApp
• Who could use the software via Citrix XenApp (i.e. has access rights)

This information gets reconciled against the entitlement data from purchase orders and software agreements.  Organizations can determine whether an additional license is needed for each Citrix user or whether the license allocated to the device for the local copy is going to cover the virtual usage. The application usage data can also tell them which users have access to the software in Citrix but don’t ever use it.  That way, administrators can tighten up access through ACL permissions on the Citrix servers.

With the Citrix XenApp, Edgesight and end-point inventory data, the license optimization tool can be used to accurately manage virtualized application licenses.  These tools also apply license entitlements—product use rights, including virtual use, to ensure that license consumption is minimized and no unnecessary licenses are purchased. 

What is the result for customers?

The customer mentioned above was able to uncover license entitlements that had previously not been utilized—meaning they had bought more than enough licenses for their current needs. They also discovered other situations where the additional licenses required were much fewer than they expected.  For some titles, they needed 1000 less licenses than they had thought.

They were also able to tighten up access to applications and are using automation to only grant access to users in Citrix XenApp where they require it.  This has resulted in significant cost avoidance and cost savings on licenses for virtualized applications.

How are you managing licenses for your virtualized applications?

***

To learn more about optimized license management for complex datacenter environments, please download our whitepaper.

It’s good to see Microsoft announcing support for the ISO/IEC 19770-2 software tagging standard. Providing and leveraging these tags comes within a six step process, where both software publishers and Software Asset Management (SAM) / Enterprise License Optimization tool providers must participate for customers to derive value. Here are the steps to tag utilization as part of the license optimization process:

 Step1: The software needs to be tagged by the software publisher. (See Flexera Software’s announcement of  end-to-end support for tagging including InstallShield 2012 support for the creation of tags and AdminStudio support for preserving tags during the repackaging process).

Step 2: The tagged software is deployed (using tools such as SCCM or Intune) to the endpoint (computer).

 Step 3: The auto-discovery / inventory tool needs to be able to collect these tags from each of the devices (desktops, laptops, servers, pads…) in the IT environment.

Today, SCCM 2012 doesn’t yet provide this tag inventory capability. It will be important for Microsoft to synchronize the future release of SCCM that is “tag-ready” and the tagging of its many software products. One without the other creates a break in the chain. Likewise, no customer value is achieved if software tags can’t be collected and processed downstream.

SAM and Enterprise License Optimization tools pick-up where SCCM leaves off. There are several more steps in license management and optimization that lead to much higher value and return on investment. Let’s take a look at the Rest of the Story, as Paul Harvey said so many times.

Step 4: Once the software tag is collected during the inventory process (future version of SCCM or other existing inventory tools—note that FlexNet Manager Platform supports tag collection now), this evidence still needs to be recognized. The capabilities required here go beyond SCCM and a SAM or License Optimization solution is required.

Tags provide richer evidence compared to other types like Add/Remove programs, but still represent fingerprints that must be processed and matched to an application recognition library entry. This recognition process produces a list of software titles per device. Now you ‘know what you have’ in your IT environment. This is the value of tagging—easing the process of accurately identifying installed software.

Step 5: A License Optimization solution (a few SAM tools provide this too) should automatically link the reconciled software titles per device to a proper set of software stocking keeping unit (SKU) numbers, where one title may have many SKUs, depending on how it was bought. The SKU number defines exactly what was purchased and how it was purchased (e.g. under an Enterprise Agreement, Select Agreement, etc.) The linkage between the title and purchase order SKUs is the key to streamlining the license reconciliation process and reporting on what was purchased compared to what was found through inventory.

This type of ‘purchased versus inventory’ report is valuable and what many companies use to defend themselves in a software license audit. However, it doesn’t account for the license entitlements that come with the software. This means that although the report is accurate from a license compliance perspective, it’s completely non-optimized, with no product use rights taken into account. In this case, you could end up paying unnecessarily for more licenses in both an out-of-compliance (audit) situation and for a new purchase scenario.

Step 6: This step is a big one, only handled by License Optimization Solutions, and disproportionate to the first five steps. It is the key to driving down spend and forestalling unnecessary software purchases. The type of agreement under which the software was purchased, the license metric/model that the vendor uses to calculate price, the software product use rights (entitlements) under the agreement, and order of license allocation are all tied together from a license/spend optimization standpoint.

Example Agreements/Terms

Microsoft:

• Enterprise Agreement
• Select Plus
• Open Value

Adobe:

•  Cumulative Licensing Program
• Transactional Licensing Program
• Enterprise Agreement

Example License Metrics

IBM:

• Processor Value Unit
• User Value Unit
• Resource Value Unit
• Authorized User
• Concurrent User
• Floating User
• Client Device

SAP:

• Developer User
• Business Expert User
• Professional User
• Limited Professional
• Business Information User
• Employee User
• Employee Self Service
• Business Analyst User

Example Product Use Rights (License Entitlements)

• Upgrade
• Downgrade
• Second Use
• Multiple Install
• Virtual Environment Use
• Reimaging
• Backup/ Fail Over/ Disaster Recovery

Example Order of License Allocation

• Standard Licenses before Professional
• Product Licenses before Suite Licenses

The license optimization solution determines the least cost outcome considering all of these variables and conditions, which are very complex and differ from vendor to vendor, physical versus virtual and even software title to title.

It’s very common for companies that reach this step to find that they are over licensed for software from some of their key vendors. They can either forestall future purchases or negotiate with their vendor and perhaps pay maintenance for only what they really need. In an software license audit scenario, the liability calculated from a software purchased versus inventory report may be greatly reduced or eliminated, after applying product use rights and generating an Actual License Position.

In March, Microsoft released new documents, the “SQL Server 2012 Licensing Quick Reference Guide” and the “SQL Server 2012 Licensing Datasheet and FAQ”. Both are available online from this Microsoft SQL Server licensing page. The most noticeable change from the previous edition is the introduction of a core factor table similar to the Oracle Processor Core factor and the IBM Processor Value Unit tables. The new Microsoft table is very simple and considers several case scenarios:

• All cores of a physical processor (socket) must be licensed with a minimum of 4 cores. This explains why the factor is 4 for single core and 2 for dual core processors.
• Some AMD processors have a core factor of .75 as they are deemed less powerful than Intel processors with an equal number of cores.

The March 2012 table is published online and is subject to change overtime. These changes will likely be driven by the release of new processor models. This is another piece of the licensing agreement that is now published online, like the product use rights document that is updated on a quarterly basis. The calculation of a license position must consider these rules, so monitoring any change is required to accurately measure compliance.

Using the Microsoft core table for physical machines is very simple. The number of cores for all processors is counted, then the core factor is applied and the result is the number of licenses that need to be purchased. For instance, if a server is equipped with 2 Intel Xeon 6 core processors, then 12 core licenses are needed as the core factor is “1” for Intel processors with 6 cores. When it comes to purchasing, core licenses are only sold by packs of two, so only 6 “packs of 2 cores” are needed. With AMD processors, the calculation may not provide an even number. For instance, an AMD FX-6100 6 core processor will require 5 core licenses (6 *.75= 4.5 rounded to 5) and the purchase of 3 “packs of 2 cores”. The extra unused core can be reallocated to another machine. Note that hyper threading is never considered for physical machines.

On virtual machines, the core table is never used. What is counted is the number of virtual cores (v-cores) supporting each virtual operating system environment (OSE), with a minimum of 4 v-cores per environment. The definition of a v-core is done at the hardware thread level: if multiple hardware threads are supporting a single virtual core, then additional licenses are needed for each thread. With the Standard Edition, each VM needs to be licensed individually. With the Enterprise Edition, there is also the possibility to license all the cores on a physical server and then run an unlimited number of SQL Server instances in physical and virtual machines, up to the number of core licenses assigned to the server. With Software Assurance, and only for the duration of it, this last limit is lifted and the number of v-cores where SQL Server is running may exceed the total number of physical cores.

Details are provided on how to migrate licenses from the processor to the core model under Software Assurance. No change is required until the end of your current agreements but the possibility exists to use the SQL Server 2012 core license model for specific deployments with new or existing licenses. At the end of their current enrolment, customers must perform and record an inventory of their servers. Their current processor licenses will be translated to core licenses according to their findings. Microsoft also guarantees a minimum number of core licenses per processor depending on edition: 8 per SQL Server Datacenter edition processor licenses and 4 for any other edition. There are two additional restrictions:

• To exchange a SQL Server processor license for core licenses, SA coverage must be renewed based on the number of core licenses required to license all of the physical cores in the processor.
• To be eligible for more than the minimum exchange, the total number of processor licenses assigned to a given server cannot exceed the total number of physical processors in the server.

The primary impact of these changes is on SQL Server licensing costs. There is no benefit in licensing old servers with less than 4 cores per processor considering that 4 cores at a minimum must be licensed for each processor. Configurations where processors have 4 cores will not be impacted by the change. Licensing SQL Server on high end processors may be significantly more expensive.

If we compare the price of SQL Server 2008 R2 Enterprise edition ($27,495 per processor) and SQL Server 2012 R2 Enterprise Edition ($6,874 per core), and apply the .75 core factor to AMD processors over 4 cores, the figures go as follows for a single processor:

As you can see, the use of newer processors with a large number of cores substantially increases the SQL Server core-based license cost.

***

To learn more about optimized license management for the datacenter, please download our whitepaper.

Software asset management (SAM) projects typically originate in traditional corporate IT organizations, whether at a global or large subsidiary level.  This makes sense, because the software publishers with highest visibility across the organization are the ones who fit squarely into the operations of central IT.  Almost everyone in the organization, including executive management, is familiar with and uses the products of Microsoft, Adobe and other major desktop software vendors.  Central IT also has ownership of IT infrastructure, so they know the importance of server-based software vendors such as IBM, Oracle and SAP.  Most critically, all these vendors, whether for general-purpose desktop software or corporate IT infrastructure software, typically impact the general corporate IT budget itself.  This means that a lack of license optimization affects the CIO’s budget.  Because of this, SAM projects usually begin by focusing on large traditional IT vendors. 

This focus on central IT scope and budget can obscure a critical area of the business that affects both software spend and software risk:  R&D and engineering organizations.  R&D organizations often have separate sourcing, purchasing, and IT processes from the main enterprise, so software asset management projects originating in central IT frequently lack visibility into software deployments and software purchases in R&D.  In addition, R&D environments often include systems that may be critical to R&D activities but fall outside the scope of central IT.  Central IT may only support Windows systems, for example, but R&D might require large Linux systems to support heavy-duty simulation work.  Finally, there can often be political tension between R&D organizations and central IT.  The potential rift between enterprise IT goals and R&D requirements can result in a siege mentality among R&D management and R&D IT staff. This can lead, in turn, to a lack of cooperation with central IT initiatives including centralized software asset management programs. 

Because R&D is often a relatively small part of an enterprise, SAM projects sometimes assume that the technical and political obstacles to including R&D environments in software asset management activities justify leaving R&D out of scope.  However, software spend in R&D is typically out of proportion to headcount, because engineering software is very expensive compared to standard desktop applications.  In fact, annual software license and maintenance spend on engineering applications can actually exceed the overall spend for software licenses and maintenance on the enterprise’s standard Windows desktop, even when R&D users are a small minority of overall users! 

In addition, R&D environments are filled with engineers.  I began my career as an R&D engineer and continue to support engineering IT teams and license management activities, and I can confidently say that engineers, as a class, possess two traits that should concern anyone interested in enterprise software license compliance.  First, engineers have the technical capabilities to acquire, install, configure, and run software in their environment.  This includes software authorized by the enterprise or by R&D management, as well as software introduced into the environment solely by the individual engineer.  Engineers also have the ability to reconfigure the software to point to a different license server than the one authorized by R&D IT, as well as remotely access systems in other geographies that may have licenses available. 

This might be manageable except for the second trait:  ignorance of the implications of software licensing on their software installation and configuration decisions.  Individual engineers are generally insulated from software contracts and purchasing processes.  As a result, they are frequently not sensitized to the financial implications or compliance and audit risks of software they have acquired and installed (through various means, some legitimate and some not).  Even in cases where an organization has an agreement with a software publisher, engineers often think in terms of site-wide licensing and may not consider that their individual actions can affect overall budget. 

The unique characteristics and risks of R&D organizations warrant their inclusion in enterprise-wide SAM projects.  While there are obstacles to such inclusion, the gains to the overall enterprise in terms of financial savings and license compliance risk mitigation mean that the effort will be worth it. 

***

To learn more about concurrent license management for engineering applications, please view our on-demand webinar: Pure Gold-- Leveraging Software Usage Data to Reduce License Costs and Risk