Manage Learn to apply best practices and optimize your operations.

How many admins do we need?

As an IT manager focused exclusively on storage, I am having a hard time proving our need for personnel. Because storage teams are a relatively new concept, the other managers do not have any experience allocating resources to a dedicate enterprise-wide storage team. Your article How many admins does it take to manage a terabyte? addresses a question regarding the ratio between personnel and storage. I concur with your comments, as I have read everything from a couple of people for 20 TB to an average of 60 TB per admin -- all of which conclude it really depends on the roles and responsibilities of the storage admins.

Currently, we have nearly 300 TB (raw) in our environment, which spans three separate sites. Our topology is site independent core/edge redundant fabrics (6 fabrics in all). We manage the disk subsystems, storage-related software, and all Fibre Channel infrastructure, plus provide reports on all the hardware including servers using detached storage on a daily/weekly basis. There are two of us on my team. We spend many late days and weekends keeping up. Our employers are discussing making a disaster recover site, and duplicating about a third of our storage environment to a forth site.

Here is my problem: I now have an opportunity to ask for personnel, but need something to reference (other than our time cards!) to justify opening positions. As an expert in storage, what are your thoughts on my situation? How many people would you expect to see in an environment like ours performing the roles of storage admin?

Thank you for the opportunity to address this important question. What you are looking at is how to justify the criticality of the needs of the storage management team so you can obtain the right level of resources.

I think the main issue that most storage managers run into is elevating the role of the storage management team within the IT organization. There is an interesting corollary to this discussion that I would like to share. Not long ago, I was visiting with the Advanced Technology Group at North Carolina State University, and I asked the assembled students if any of them had taken courses in managing storage or storage technologies. None of them raised their hands. I then asked if any of the assembled students took classes in host processing technologies. Every student raised his or her hand. I was surprised to find that the university, a major information technology and engineering powerhouse, doesn't offer classes focused on storage technologies. No SCSI, no Fibre Channel, no volume manager theory, no file-system theory, no designing storage environments for higher availability, and so on.

My point is that the importance of storage in the overall solution is often overlooked. Part of the problem is that the IT organization expects storage to just work. Also, in the past, there were not any choices amongst the tiers of storage. Everyone in the organization either got a local disk in their host for their application, or their host was connected to a large disk array via SCSI or FCP. Replication occurred through mirroring and everyone backed up to a local tape array.

Now storage has evolved. There are tiers of storage providing a full spectrum of offerings from high availability and remote DR to low cost local disk arrays with SATA drives. The IT organization needs to be involved with the planning process for a new project all the way through the management and archiving of data. This is a new way to think of the role of the storage management group and the storage manager (yourself) needs to engage in a marketing effort on behalf of the group to elevate the challenges within the storage group to the rest of the organization's business process owners.

You'll have to do some pre-work to document what the group does and what the group is looking to do in the near future. I would suggest that you look at pulling together some real numbers based on the following:

  1. Storage pool size: A historic and future look
  2. Change rate of data growth, based on:
    • Current headcount
    • Number of current projects and projected growth of projects
    • Costs for the efforts
  3. Hardware and software acquisition costs
  4. Host and host software costs
  5. Number of help desk tickets opened for storage and cost per ticket.
This process should yield a better understanding of the way things work today. After that, you should set clear goals going forward, and review your progress at set times. I recommend checkpoints of 30 days, 60 days, 90 days, 180 days, one year, eighteen months and two years. This should all be balanced against the company's goals and demands.

My hope is that you aren't already so underwater or stressed with the current work efforts in the storage management arena that you can look at measurable goals in the form of SLAs for the resources that you have today and will get in the future.

Let me know your thoughts and concerns.


Brett P. Cooper

Dig Deeper on Data storage strategy

Start the conversation

Send me notifications when other members comment.

Please create a username to comment.