HP MSA1000 storage works raid array 201723-B21 전단
제품 코드
201723-B21
© 2002-2006 Data Mobility Group. All Rights Reserved. 76 Northeastern Blvd. Suite 29A, Nashua NH 03062 Phone: 603.835.6141 Fax: 877.254.4886
Data
Mobility
Group
www.datamobilitygroup.com
Tiered Storage, Tiered Data, Unified Management
Historically, significant savings can be realized from minimizing capacity growth on the gold-standard
disk arrays by moving data off to higher capacity, lower cost-per-GB storage devices. Today, the
challenge is to be able to do so in such a way that:
disk arrays by moving data off to higher capacity, lower cost-per-GB storage devices. Today, the
challenge is to be able to do so in such a way that:
• Application performance is not negatively impacted. In fact, performance gains should be
expected from moving less frequently accessed data off of high-performance disk arrays.
• Management complexity is not magnified by creating multiple, dis-integrated management
consoles.
• Unnecessarily redundant copies of data are not created. Rather, one of the goals of a storage
tiering project should be to reduce the number of data copies.
• Data is not left in an unprotected - or perhaps worse - in an unsecured state as may be the case
with backup tapes stored in remote offices.
• Retrieval of data that must be promoted back to a higher tier as required by an application or
application user is not hampered.
Early storage tiering implementations exposed these issues, suggesting the dual need for a data
classification scheme coupled with some way to move data non-disruptively up and down the physical
storage tiers over time. This realization created a critical need for unified storage management
software. It is not enough to partition the storage environment into “buckets” ranging from high- to
intermediate- to low-performance and availability. One must also have an integrated view of the tiered
infrastructure as well as some way to classify data and automate the allocation of data over time based
on classification.
classification scheme coupled with some way to move data non-disruptively up and down the physical
storage tiers over time. This realization created a critical need for unified storage management
software. It is not enough to partition the storage environment into “buckets” ranging from high- to
intermediate- to low-performance and availability. One must also have an integrated view of the tiered
infrastructure as well as some way to classify data and automate the allocation of data over time based
on classification.
To do this, storage administrators need centralized data management views and functions layered on
top of the physical tiered storage architecture. This may require vendors to move some data and content
management functionality that has traditionally lived at the application level down into the storage
infrastructure as well as introduce tighter integration between performance monitoring and migration
so that data sets can be temporarily demoted to a lower physical tier or promoted to a higher tier as
production processing and recovery needs dictate.
top of the physical tiered storage architecture. This may require vendors to move some data and content
management functionality that has traditionally lived at the application level down into the storage
infrastructure as well as introduce tighter integration between performance monitoring and migration
so that data sets can be temporarily demoted to a lower physical tier or promoted to a higher tier as
production processing and recovery needs dictate.