Today we continue discussing the best practice from storage engineering:
473) Storage products are also prone to increasing their test matrix with new devices such as solid state drive and emerging trends such as IoT
474) Storage products have to be limitless for their customers but they cannot say how they will be used. They will frequently run into usages where customers use them inappropriately and go against their internal limits such as the number of policies that can be applied to their organizational units.
475) There was a time when content addressable storage was popular. It involved generating a PEA file to save contents that could be looked up by their hash. The use of object storage made it easier to access the objects directly.
476) Data is increasingly being produced as fixed content Emails and faxes are examples of these. The lifecycle of content such as from system, personal computing, Network centric and content centric are progressively higher and higher in their durations
477) Drop and create of user artifacts helps the user to cleanup. This is not the case for say system catalog. Still the storage artifacts used on behalf of the user is also the same as the storage artifacts used for system itself. Creating and dropping such artifacts would be helpful even if they are internal.
478) The retention policy is typically 6 months for email, 3 years for financial data, 5 years for legal. The retention period for object storage is user defined.
479) Object Storage is touted as best for static content. Data that changes often is then said to be preferred in NoSQL or other unstructured storage. With object versioning, API and SDK, this is no longer the case.
480) Data Transfers have never been considered a virtual storage since they belong to the source. Data in transit can live in queues, cache and object storage which is good for vectorized execution .
473) Storage products are also prone to increasing their test matrix with new devices such as solid state drive and emerging trends such as IoT
474) Storage products have to be limitless for their customers but they cannot say how they will be used. They will frequently run into usages where customers use them inappropriately and go against their internal limits such as the number of policies that can be applied to their organizational units.
475) There was a time when content addressable storage was popular. It involved generating a PEA file to save contents that could be looked up by their hash. The use of object storage made it easier to access the objects directly.
476) Data is increasingly being produced as fixed content Emails and faxes are examples of these. The lifecycle of content such as from system, personal computing, Network centric and content centric are progressively higher and higher in their durations
477) Drop and create of user artifacts helps the user to cleanup. This is not the case for say system catalog. Still the storage artifacts used on behalf of the user is also the same as the storage artifacts used for system itself. Creating and dropping such artifacts would be helpful even if they are internal.
478) The retention policy is typically 6 months for email, 3 years for financial data, 5 years for legal. The retention period for object storage is user defined.
479) Object Storage is touted as best for static content. Data that changes often is then said to be preferred in NoSQL or other unstructured storage. With object versioning, API and SDK, this is no longer the case.
480) Data Transfers have never been considered a virtual storage since they belong to the source. Data in transit can live in queues, cache and object storage which is good for vectorized execution .
No comments:
Post a Comment