HP Peer Motion Makes Being Wrong Alright
No one ever gets every decision right the first time. In fact, most of us are lucky to get most of our decisions mostly right most of the time. Yet to date storage administrators have been forced to live in somewhat of an alternate universe where the expectation is that they plan and execute on storage allocations perfectly every time.
To say that storage administrators must live up to unrealistic expectations is an understatement. In their world everyone expects them to always be right when provisioning storage and for that decision to be correct months or even years later. The good news for them is that the new HP Peer Motion feature on HP 3PAR and HP LeftHand Storage Systems removes some of that pressure as they can now occasionally be wrong and still be alright.
Anyone who has ever worked as a storage administrator knows that when they receive a request for storage capacity, the application owners only know that they need X number of GBs or TBs of storage capacity.
These application owners are not necessarily thinking about the cost of the underlying disk drives, their performance characteristics, or what protocols are optimal or most cost effective for the application. They just want storage capacity and the storage administrator to deliver it so they can keep their project moving forward.
This leaves storage administrators in a bit of quandary. Either they are forced to ask a lot of questions that the application owner may not be ready to answer such as:
Yet neither of these options is really optimal from a business perspective. The first option may take too long while straining relations between the application owner and storage administrator if the application owner is unprepared to answer questions about what type of storage their application needs. Conversely, if a storage administrator just provisions whatever storage is available to the application even though it may be too expensive, under perform or both.
Now a normal person would just say, "Well, if the application is using a volume that resides on the wrong tier of storage, just move it to the right tier." To which a storage administrator may lament in response, "If it was only that easy."
Any savvy storage administrator knows that moving a volume from one tier of storage to another, much less from one storage array to another, could take days, weeks or even months to execute upon. Such a move may entail application outages, server reboots, changing LUN security, SAN switch zoning changes, replication software and even HBA firmware updates.
This pain of making such a move illustrates why storage administrators almost always feel such pressure to provision storage perfectly the first time every time. Further, it also highlights why storage administrators prefer to use the most highly available, best performing (and typically most expensive) storage. They do not want to come back and re-provision the storage later if the array should it under perform or run out of storage capacity.
This decades-long issue is what storage federation technologies like the new HP Peer Motion feature on HP 3PAR and HP Lefthand storage system addresses. Storage federation enables native controller based communication between like storage systems so volumes are managed at a data center or campus level rather than only at the individual device level.
HP Peer Motion is a federated capability that enables administrators to seamlessly move a volume from one storage array to another that has the extra capacity and extra performance that the volume may need without all of the LUN security, zoning changes, firmware updates and replication software that are all needed in part or in whole to execute upon such a move.
This can also work in reverse in cases where a volume may not need the performance or capacity it currently has so it can now be moved to another storage array that has less of these features to offer.
It is this aspect of being able to move a volume from one specific storage array to another as opposed to just moving a volume from one set of disks to another within a storage array that is arguably the most intriguing and exciting aspect about Peer Motion. This creates opportunities for organizations to optimize the utilization of storage arrays across their data center or campus as opposed to just being able to optimize storage utilization within an individual storage array.
The illusion that storage administrators always get storage provisioning decisions correct is just that: an illusion. Administrators are human beings too and cannot and should not be expected to "know all and always be right" in regards to perfectly placing volumes in the right place every time. But to date they have been boxed in by inflexible technologies that preclude them from easily moving volumes after it has been provisioned.
HP Peer Motion on the HP 3PAR and HP Lefthand systems break this barrier - probably much to the relief of both storage administrators and the organizations for which work. By giving storage administrators new flexibility to move volumes on as needed basis where it is best for the application and the business with minimal effort on the part of the storage administrator, they can provision storage more quickly without worry, applications can grow as needed and everyone wins - which is alright.
To say that storage administrators must live up to unrealistic expectations is an understatement. In their world everyone expects them to always be right when provisioning storage and for that decision to be correct months or even years later. The good news for them is that the new HP Peer Motion feature on HP 3PAR and HP LeftHand Storage Systems removes some of that pressure as they can now occasionally be wrong and still be alright.
Anyone who has ever worked as a storage administrator knows that when they receive a request for storage capacity, the application owners only know that they need X number of GBs or TBs of storage capacity.
These application owners are not necessarily thinking about the cost of the underlying disk drives, their performance characteristics, or what protocols are optimal or most cost effective for the application. They just want storage capacity and the storage administrator to deliver it so they can keep their project moving forward.
This leaves storage administrators in a bit of quandary. Either they are forced to ask a lot of questions that the application owner may not be ready to answer such as:
- What type of performance does the application need?
- What type of availability does it need?
- How fast will the application grow?
Yet neither of these options is really optimal from a business perspective. The first option may take too long while straining relations between the application owner and storage administrator if the application owner is unprepared to answer questions about what type of storage their application needs. Conversely, if a storage administrator just provisions whatever storage is available to the application even though it may be too expensive, under perform or both.
Now a normal person would just say, "Well, if the application is using a volume that resides on the wrong tier of storage, just move it to the right tier." To which a storage administrator may lament in response, "If it was only that easy."
Any savvy storage administrator knows that moving a volume from one tier of storage to another, much less from one storage array to another, could take days, weeks or even months to execute upon. Such a move may entail application outages, server reboots, changing LUN security, SAN switch zoning changes, replication software and even HBA firmware updates.
This pain of making such a move illustrates why storage administrators almost always feel such pressure to provision storage perfectly the first time every time. Further, it also highlights why storage administrators prefer to use the most highly available, best performing (and typically most expensive) storage. They do not want to come back and re-provision the storage later if the array should it under perform or run out of storage capacity.
This decades-long issue is what storage federation technologies like the new HP Peer Motion feature on HP 3PAR and HP Lefthand storage system addresses. Storage federation enables native controller based communication between like storage systems so volumes are managed at a data center or campus level rather than only at the individual device level.
HP Peer Motion is a federated capability that enables administrators to seamlessly move a volume from one storage array to another that has the extra capacity and extra performance that the volume may need without all of the LUN security, zoning changes, firmware updates and replication software that are all needed in part or in whole to execute upon such a move.
This can also work in reverse in cases where a volume may not need the performance or capacity it currently has so it can now be moved to another storage array that has less of these features to offer.
It is this aspect of being able to move a volume from one specific storage array to another as opposed to just moving a volume from one set of disks to another within a storage array that is arguably the most intriguing and exciting aspect about Peer Motion. This creates opportunities for organizations to optimize the utilization of storage arrays across their data center or campus as opposed to just being able to optimize storage utilization within an individual storage array.
The illusion that storage administrators always get storage provisioning decisions correct is just that: an illusion. Administrators are human beings too and cannot and should not be expected to "know all and always be right" in regards to perfectly placing volumes in the right place every time. But to date they have been boxed in by inflexible technologies that preclude them from easily moving volumes after it has been provisioned.
HP Peer Motion on the HP 3PAR and HP Lefthand systems break this barrier - probably much to the relief of both storage administrators and the organizations for which work. By giving storage administrators new flexibility to move volumes on as needed basis where it is best for the application and the business with minimal effort on the part of the storage administrator, they can provision storage more quickly without worry, applications can grow as needed and everyone wins - which is alright.
Newsletter Signup
Leave a comment