|
|
|
|
 |
| Tech Update
|
|
Add servers, not admins
Improve staffing ratios
|
By Brian Richardson
February 11, 2002
| Provided by |  |
|
|
|
Q: Why are there two ratios (in other words, end user-to-sys-admin or server-to-sys admin)?
A: Just as there is a wide range of sys admin job descriptions and daily activities, there is also a significant difference between sys admin roles in primarily file and print, mail, or PC-LAN server environments and in those that support more complex application infrastructures (for example, for ERP, CRM, data warehouse, and e-business). For PC-LAN environments, we recommend that IT organizations focus on the ratio of end users to sys admins, because most TCO and support costs are centered on end-user issues. The three primary systems administrator activities in these cases are office or other personal application support, user move/add/change (MAC) requests, and break-fix maintenance support. For application server or DBMS environments, however, we believe the relevant ratio is the number of servers to systems administrators, because sys admin responsibilities are largely independent of variations in the actual number of users (in other words, the number of applications is key). For file and print environments where sys admin support primarily includes MAC and break-fix support (but not application support), the server-to-sys admin ratio can be combined with the application/server ratio.
Q: OK, but what is a best-practice guideline?
A: Our research indicates that most IT organizations have a system-to-sys admin range of between 10:1 and 20:1. Accordingly, we believe shops with a server-to-sys admin ratio greater than 30:1 can make a good case for additional sys admin staff. Yet IT organizations with a ration of less than 10:1 may be overstaffed. Some estimated server-to-sys admin range variability is allowed. In addition, for very large IT organizations, one or more layers of systems administrator management must be included. For example, in a shop with 1,500 servers, nominally 100 sys admins would be appropriate. However, an additional 12 to 15 sys admin management positions must also be included (assuming about eight systems administrators per manager, plus second-level management). Moreover, our research indicates that, in most IT organizations, infrastructure consolidation initiatives do not generally result in staff reductions. Instead, the primary goal is to increase the number of applications (and attendant infrastructure), without requiring lock-step headcount growth.
Q: What initiatives improve staffing ratios?
A: We believe the most important first step in improving staffing ratios is physical co-location from a large number of sites to a small number of regional data centers. The next step is evaluation of storage-area networks (SANs). However, it is important to note that SANs are not a silver-bullet technology. Without mature storage management polices and procedures, SANs will introduce additional complexity and increase (rather than decrease) storage-related costs. Restricting the number of different server configurations also reduces complexity; supporting, for example, only small, medium-size, and large systems, with nearly identical configurations in each category. In addition, reducing the number of different versions of software (operating system, DBMS, application, etc.) tends to reduce complexity and costs.
Our research does indicate a few rare cases with server-to-sys admin ratios greater than 50:1. We believe these are relatively unique situations that should not be viewed as general best-practice goals. All of the IT organizations with ratios greater than 50:1 exhibited several common characteristics. In each case, identical server systems were being deployed in a large number of remote or branch locations. In addition, only a small number of in-house developed applications (typically one) were being used, and a significant investment had been made in custom management and reporting software. In addition, all of these shops had already developed mature operational best practices, in some cases running several thousand servers.
The Bell Curve: Server-to-System Administrator Ratios
First published on January 8, 2002
By Brian Richardson
Visit Meta Group for more IT research and analysis.
 |
|
|
|
[an error occurred while processing this directive]
![]() |
|
[an error occurred while processing this directive] |
![]() |
|
![]() |
[an error occurred while processing this directive]
|
|
[an error occurred while processing this directive]
|
[an error occurred while processing this directive] |
|
|
|
|
|
