The Storage Networking Industry Association (SNIA) recently surveyed a number of end users on their main concerns with storage. These are the top five, in order of priority. There aren’t any surprises here.
Tiered storage, as part Application Optimized Storage Solutions, addresses these pain points head on. But in order to understand the specific role tiered storage plays, we need to look at what Hitachi Data Systems means by Application Optimized Storage solutions.
The key message is that HDS is creating solutions that enable organizations to optimize storage infrastructure based upon application requirements. It is important to differentiate between optimizing applications with storage vs. optimizing storage to match application requirements. HDS does not claim to optimize applications, as that falls outside the core competencies of a storage company.
Our core competency is optimizing storage based upon application requirements, in support of top line business objectives.
The key message is that HDS is creating solutions that enable organizations to optimize storage infrastructure based upon application requirements. It is important to differentiate between optimizing applications with storage vs. optimizing storage to match application requirements. HDS does not claim to optimize applications, as that falls outside the core competencies of a storage company.
Our core competency is optimizing storage based upon application requirements, in support of top line business objectives.
If we look at a typical enterprise system we would find data stored on potentially inappropriate hardware, the graph here shows data storage in decreasing order of reliability (and therefore value) and data in terms of organisational value, maximum at the top {CLICK} In most systems today some of the important Data is in the most appropriate storage. In this example we have the Billing and Data Mining Data on the current most sophisticated solution, the Lightning 9980v - but the ERP system, despite being potentially the most important current data asset, is on the least appropriate system, either because it was the first system to be placed in the SAN or because of the complexities of moving it elsewhere. By introducing USP {CLICK}
We can use the USP’s management system to move the high value data immediately onto the fastest most reliable storage hardware in the system {CLICK} but, perhaps more importantly, we can tell the USP to over time realign the complete data centre against the most appropriate storage solution {CLICK}. Thus USP allows the experienced Data manager to move for the first time towards a Tiered Storage Solution, vastly improving the SAN performance and maximising the return on this considerable corporate asset {CLICK}
The Storage Networking Industry Association defines virtualization as:
“An act of integrating one or more (back end) services or functions with additional (front end) functionality for the purpose of providing useful abstractions. Typically virtualization hides some of the back end complexity, or adds or integrates new functionality with existing back end services.”
Okay, maybe a bit clinical, but it does highlight the point that virtualization is suppose to hide some of the complexity of storage. It is a key enabler to implementing and managing tiered storage.
Virtualization overall:
Allows for all multidimensional tiers to be consolidated into one pool
Enables management of storage as one pool rather than separate tiers from a central console
Enables such tasks as adjusting tiers and moving data without affecting the SAN, hosts, or applications
Simplifies the implementation and management of storage tiers
The Universal Storage Platform and Network Storage Controller offer a unique implementation of virtualization within the storage array. Not only does the Universal Storage Platform and Network Storage Controller offer virtualization, but it has a unique capability to logically attach additional storage arrays and virtualize those volumes and present them to the SAN as one storage pool, both to Open Systems hosts via the SAN and to mainframe hosts.
To understand the need for tiered storage, customers should know the problems this solution will solve.
Hitachi Virtual Partition Manager
Logical partitioning’s main enhancement is the partitioning of cache which is required to provide QoS. It prevents a heavy cache user from impacting other users who share the same TagmaStore Universal Storage Platform. With the Lightning 9900 V Series and Thunder 9500 V Series we could guarantee safe multi tenancy with Host Storage Domains. Multi tenancy means that multiple users can share the same storage port without over writing or seeing each other’s address space. Multi tenancy alone can not guarantee QoS unless the cache is also partitioned.
This allows mission critical applications to share the same TagmaStore Universal Storage Platform. Critical applications used to require their own storage systems, since you would not want one application to do a data base re org while another critical application is going through peak processing. The logical partitions also have virtual serial numbers for chargeback.