Storage Management Strategies for the CIO
Let's look closer at real problems that administrators have to deal with that consolidation solves:
How do you back up hundreds of servers?
Requires Free Membership to View
Very slowly, even if you have a utility segment for backing up data. TCP/IP data even over GigE is not an efficient way to backup large amounts of information. Large corporations and agencies often have 50, 100, 500, or many times more servers in their data centers, often in racks of independent systems. Often they each have their own storage, console keyboard and display-independent of all other systems. Management of this environment is becoming increasingly complex. In many environments, servers multiply to meet the needs of distributed computing. As implementation grows, more and more servers are added, each one requiring extra RAID disks -- and spares -- and individual backup. Pretty soon the task of managing this environment can become challenging. Critical data can be lost -- or huge amounts of resources may be required to perform the daily backups in a world of ever shrinking backup windows. Having a consolidated data infrastructure means that the backups can be performed off of the LAN and directly from the high performance SAN infrastructure. As one example, using virtualization (either from tools within a vendors storage array solution or using in-band or out-of-band virtualization software and appliances) you can achieve 100GB/Hour backups AND restores per LTO drive and block level volume backups. If application agents are used, backups can be accomplished with no impact on the production access of data.
How do you insure efficient utilization? With separate storage on each server none of the sharing and protection is available that consolidated storage offers. There are no economies of scale. If one server needs additional disk, it is of no help that another server is underutilized. Sometimes -- in order to address hundreds of GB of storage, it is necessary to add a server -- just to obtain the slot space to connect to the required disk capacity. The drawing below illustrates an example of how utilization between servers with different kinds of RAID storage can vary. The cylinders represent different types of server platforms and a typical pattern of utilization. Without a consolidated storage pool, unused space from one type of platform could not be allocated to another. Data consolidation can also provide economies of asset utilization not available with standalone servers. If disks can be allocated dynamically, servers get only the storage they need. Utilization efficiencies of disk space goes way up. And spares are available to all the servers on the system.
How do you reduce storage management costs? According to a study performed by Cahners In-Stat Group, the cost of managing storage has risen to 91% of the cost of the media, or about $3.50 per year per megabyte. While this trend is staggering, what's even more troubling is the fact that most companies are running at 40% utilization--for every $1 spent on storage assets, $0.60 is wasted.
Editor's note: Do you agree with this expert's response? If you have more to share, post it in one of our .bphAaR2qhqA^0@/searchstorage>discussion forums.
This was first published in October 2002
Join the conversationComment
Share
Comments
Results
Contribute to the conversation