Software upgrades: More gain, less pain
In this second installment of our two-part software selection guide, Gartner focuses on the issues relating to open source migration and offers tips to better handle software upgrades.
Desktop management is important to small and medium-size businesses as its associated costs are largely non-discretionary and increase in direct proportion to growth of the organization.
As the number of employees increases, so does the number of desktops. Couple this with the replacement of PCs (and increasingly larger share of laptops), desktop costs could grow rather rapidly.
The challenge is further compounded by the limited choices that IT managers have since the desktop software market is dominated by one vendor--Microsoft.
Broadly, desktop software comprises four segments--operating system, productivity suite, client-access license for server software, and desktop business applications such as accounting.
Many IT managers today are faced with two important decisions regarding desktop software issues: moving to open-source software and planning upgrades.
| When it comes to migration to open source, many organizations make the mistake of confusing the operating system with the productivity suite. |
Open source
When it comes to migration to open source, many organizations make the mistake of confusing the operating system with the productivity suite. It is common to find IT managers contemplating or executing standardization to Linux with the intent of running, say, StarOffice or OpenOffice.org, which cost significantly less. Such confusion may be a costly mistake since these productivity applications may run on their current version of Windows as well.
There is no need to migrate to Linux as the client operating system, which will bring its own challenges. Therefore the decision about the productivity suite must be separated from the client operating system.
Also, firms need to be realistic about other issues related to Linux on the desktop. It is commonly believed that Linux is free. While consumer versions are indeed free, supported versions are not.
SMBs that use these versions have to rely on the open source community for support. IT managers need to decide if the quality and support meets their needs. If not, they will have to go for vendor supported versions that come for a price, which may be lower than that paid for commercial products, but it is a cost nevertheless, and must be factored in. Further, much of the cost of migration (both from Windows to Linux and from one version of Linux to another) is people related and that may not be substantially different in the two scenarios.
It is also a myth that with Linux, upgrades are not necessary. Sure, there is no pressure from a vendor to upgrade, but support issues from the OS vendor as well as the application vendors may force users to upgrade. Not all the versions of Linux that have ever shipped are supported. Similarly, ISVs providing applications on Linux may not support them on older versions of Linux indefinitely. This should also be part of the calculation.
IT managers need to take note of the increased vendor activity in the open source space. HP is now selling some desktop models with Linux and open source productivity software. However, SMBs should look for proof points of value before taking a decision to migrate. These are not arguments against Linux, but to emphasize the need for caution when taking the decision.
Planning for upgrades
Another issue often faced by IT managers is when to upgrade the operating system and productivity suite. This is never an easy issue. Upgrades are costly and sometimes benefits are not tangible. However, it is also difficult and costly to maintain several versions of software. Also, when a version of software is not supported anymore, users cannot expect any bug-fixes, which may compromise performance and security in the future.
At the same time, upgrading involves many decisions. Should IT managers upgrade each time a new version becomes available or skip some versions? Should they pay maintenance to have rights to upgrades or license new versions when needed? Should they simultaneously upgrade software on all desktops or roll them out in phases?
Here are some guidelines that aid decision-making:
1. One way to handle upgrades is to look at the combined hardware-software refresh. Sometimes it is cheaper to buy new hardware than to support the old, as hardware costs are dropping all the time. OEM licensing of operating systems and productivity suites may be a cheaper way for organizations that do not upgrade often.
2. The decision to upgrade all desktops at once or in stages depends upon many factors. Although the former reduces the diversity and therefore management costs, it can be expensive as some supported versions will be replaced. This cost should be set off against the benefits. Also, this approach may lead to a prolonged deployment cycle, with the next replacement cycle looming large.
3. Sometimes it may be worthwhile to skip a version of a productivity suite so that an organization may aim for a single version at a future point of time. This may mean delaying upgrade on some desktops even if the organization has upgrade rights. It may also mean running with unsupported versions on some desktops until the target version upgrade. It is more difficult to defer operating system upgrade since there usually is a time gap between different versions. SMBs should keep the benefits and risks in sight when taking this decision.
4. Whether to pay for maintenance or not depends upon how frequently an organization intends to upgrade. This depends on the characteristics or needs of the user organization as well as the users’ perception of value to be delivered in the upgrades announced by the vendor. If the user organization believes that it does not need to upgrade for three-and-a-half years or more, then it may opt for fresh licensing at full cost rather than paying maintenance (called Software Assurance by Microsoft).
5. One important factor to keep in mind is heterogeneity of operating system in the IT environment. Typically, in an IT set-up with fewer than 500 desktops, which is typical of most SMBs, heterogeneity is a big TCO (total cost of ownership) multiplier, because of a lack of economies of scale. Thus, resources spent on achieving homogeneity may pay for themselves by reducing the long-term cost of maintenance.
6. The timing of upgrade may also be impacted by special offers that a vendor may release, such as option to spread the payment over a period of time, and special rebates.
SMBs should be driven by the goal of long-term TCO reduction when taking desktop software related decisions. They should objectively consider all options, such as OEM licensing, open source, vendor special offers, phased implementation and skipping some versions, before deciding what combination works best for them.
Pranav Kumar is research director for Enterprise Application Software, Gartner Asia Pacific. He is also a member of the CNETAsia SMB Advisory Board.
