The shifting landscape of the cloud in Software Asset Management: How you can survive it

In the next two years, industry experts are predicting that over half of a company's IT applications will be located in the cloud. For asset management professionals, this requires yet another shift in methodology and technology, due to both increasing complexity on the network and a shift in the way that asset management is perceived by businesses. Meanwhile, it's likely that the number of software applications needing to be managed will also increase, so it's now more important than ever to be up to date with current developments.

Traditional 'WINTEL' environments of the past have given way to heterogeneous platforms combining Linux, Mac, Windows, virtualisation, and more recently software as a service (SaaS) applications. In a cloud environment, it's also necessary to consider how your company can manage a mixed estate of public and private clouds.

In difficult cases, some applications will reside locally or in a private cloud, whereas others will be hosted in a public cloud, as a virtual desktop for instance. Analysts are predicting a significant rise in the use of mixed cloud environments like these, and it can be unclear where the responsibility for complying to licenses lies in such situations.

The challenges of licensing in a mixed environment

Having to manage such a mixed environment brings to the fore many technical and licensing challenges for enterprises to overcome. These challenges primarily fall into two categories:

1. As new platforms launch, they require users to follow their own proprietary licensing schemes. Problems can sometimes arise when these new schemes don't easily align with the traditional direction of many Software Asset Management (SAM) programmes, which provide one-to-one device or user licensing verification.

2. The new platforms themselves often require new technologies in order to accurately track the deployment and usage of both hardware and software assets. Again, traditional WINTEL solutions are fast becoming obsolete.

As an example of how licensing has become more complex, just look at a vendor such as Oracle. Its licensing models are highly complex, especially for virtualised environments. Users have to factor processing power into their calculations, as well as the level of integration with other applications and the numbers of users requiring access. For organisations that are integrating cloud-based applications with Oracle, it becomes even more difficult to manage and isolate the interrelationships between the different applications and users who access them.

Oracle's virtualisation rules can also be incredibly complex for users in an enterprise environment. Oracle doesn't support soft partitioning on VMware clusters, so it's important to have full visibility of exactly where Oracle is deployed in the estate and what the underlying virtual server architecture relates to.

A common scenario we uncover when working with customers is that the IT administrator installed Oracle database products on a virtual server, but failed to recognise the licensing impact of this decision. They may have assumed that only the virtual server needed to be licensed, when in fact they should have taken into account the underlying architecture of the physical servers.

In the case of a VMWare cluster, the whole cluster needs to be licensed for every physical server where the virtual server might be running. Not having a clear understanding of Oracle licensing can put the organisation at risk of non-compliance, and a small oversight like the one above can become a very expensive mistake – sometimes to the tune of millions of pounds in penalties.

IBM's Byzantine model

IBM is another vendor with a notoriously complex licensing model. It uses the PVU (processor value unit) metric instead of counting the numbers of users requiring access. In this instance, a users' licensing obligations (and the costs they will incur), are calculated in proportion to the relative power of the hardware installed.

For the user, having to navigate this kind of complexity makes it difficult simply to ensure compliance, and to purchase the correct volume of licenses, let alone consider whether their software licensing contracts could be negotiated more cost-effectively.

As the switch to cloud-based applications continues to gain pace, this will upset the traditional view of SAM as primarily delivering licensing compliance. This may not be such a bad thing if it forces a shift towards SAM focusing on proactive optimisation and cost-prevention. For IT asset managers, this means they have an opportunity to make a greater contribution to their organisation and demonstrate a tangible return on investment.

What is a cloud-based application, exactly?

Before discussing the reasons behind this change, let's clarify what we mean by a cloud-based application. Although definitions vary, one aspect common to all cloud software applications is that they cannot, by nature, create non-compliance problems. This is because although the exact delivery model can vary, the user purchases an entitlement to access the software, which is then controlled directly by the vendor.

For instance, Adobe's version of Creative Cloud isn't, strictly speaking, delivered via the cloud in the same way that Salesforce is. Adobe software resides locally on a device and it's the license authentication component that's managed in the cloud. Contrast this with Salesforce, which is a 'true' cloud application, since nothing gets installed on the user's device. Here we have two different models, but in both cases, compliance is no longer an issue because it's automatically taken care of by the vendor.

There's even further uncertainty surrounding what's being called the hybrid cloud, where organisations use a combination of publicly hosted applications - for instance a virtual hosted desktop with some applications hosted on their own private cloud. Analysts have suggested that this will create problems over where the responsibility for compliance lies and how licensing obligations can be calculated.

We have yet to assess the detail of exactly how this will create difficulties but can however imagine the increased complexity it will bring. For example, an organisation might use a public cloud service, such as Windows Azure Blob Storage Service for archived data (simply for the ability to scale), but continue to maintain in-house storage and software for operational purposes.

Getting value for money

Scenarios like these present very different license management challenges for an organisation. Rather than concern with compliance and audit avoidance, the emphasis switches to optimisation and cost-avoidance through prediction of licensing requirements. Because it's no longer in the interests of the software vendor to expose over-purchasing, the onus is increasingly on the user to identify whether they are getting value for money from their investment in the software.

  • Is it being well used?
  • Is there potential to review licensing agreements and reduce the numbers of seats purchased?

This is the direction SAM solutions will be heading into in the future. They will have to actively support organisations by providing them with the intelligence they need to become more cost-efficient. For the enterprise, the shift away from software asset compliance creates new opportunities.

Now businesses can precisely measure the value software investments are bringing, and minimise wastage. Users that have been hit by a compliance failure or software audit in the past will find this particularly useful, since they often err on the side of caution and over-buy their licenses. When software is accessed via the cloud, this tendency can be reversed because intelligence is automatically gathered to enable precise utilisation patterns to be identified.

We can expect software vendors to increasingly offer transaction-based licensing. Here the user organisation might purchase an entitlement for a set number of licenses over a period of time, and will be looking for the intelligence to identify whether they have correctly predicted their requirements. This will, once again, shift both the focus of SAM programmes and the requirements placed on the SAM solutions deployed.

The licensing changes described above demand an evolution in SAM tools more radical than anything the industry has seen in the last five years. Whereas once having Windows discovery was sufficient to claim activity in Software Asset Management, the mechanics of an effective SAM tool is changing extremely quickly.

For example, it will no longer be enough for agents to simply discover the applications physically installed on a desktop. They will also need to monitor the use of cloud-based applications by individual users and perhaps even start to track individual transactions rather than overall activity levels.

What next for SAM experts?

Looking ahead to the coming 18 months, SAM professionals can expect to see the capability to monitor usage for transaction-based licensing and other new capabilities becoming increasingly available. For instance, automated re-harvesting may soon be commonplace. This will take the form of a license automation agent capable of monitoring access rates to different software solutions and actively tracking usage patterns. It may even have the ability to uninstall unnecessary applications and make their spare licenses available to others with a greater need.

For example, if a user has access to Salesforce and records show that the application has only been used once in a three month period, automated re-harvesting could be set up to alert IT to the situation. Then the user would receive an email asking for confirmation of the need for the subscription. If no response was received, the application would be uninstalled automatically and the license entitlement added to an available 'pool' for another subscriber to utilise. In this way organisations can be sure that their entitlements are being used in the most efficient way and avoid incurring the extra administration costs of manually checking usage levels.

In addition, licensing logic built into the next generation of SAM solutions will help users become aware of their entitlements and ensure they get maximum value from negotiated contracts. This is valuable because it is not a requirement of the software vendor to inform users pro-actively of the terms of their licensing agreements.

For instance, Microsoft users will have the intelligence to know whether specific users are entitled to free software upgrades, e.g. for Office 2013, as part of their original licensing agreement. Alternatively, they will have the potential to run simulations to calculate exactly what their licensing obligations and costs should be for specific vendors to prevent over-purchasing.

Finally, we will see the more advanced SAM solutions become more like their Business Intelligence cousins, where planning, forecasting and 'what if' analyses become a larger part of the offering – again helping SAM managers not only to ensure compliance, but also to deliver real value back to the organisation in terms of cost-avoidance and asset optimisation.

Functionality like this will help to make software asset management a business requirement in its own right. As the shift away from policing and compliance continues to gather pace, and proactive cost control and optimisation are becoming more and more integral, SAM solutions will become the ultimate source of independent verification. Rather than simply helping to limit the financial impact of a software audit, they will allow companies to predict future consumption levels, to provide ongoing intelligence about actual usage levels, and ensure they are getting what every company wants - real value for money.

Peter Bjorkman is CTO of Snow Software and an expert in software asset management. For more information, visit www.snowsoftware.com

Images: Flickr (torkildr, brka, msaari)

Main image credit: Flcikr (masahiko)