[fusion_builder_container hundred_percent=”no” hundred_percent_height=”no” hundred_percent_height_scroll=”no” hundred_percent_height_center_content=”yes” equal_height_columns=”no” menu_anchor=”” hide_on_mobile=”small-visibility,medium-visibility,large-visibility” class=”” id=”” background_color=”#ffffff” background_image=”” background_position=”center center” background_repeat=”no-repeat” fade=”no” background_parallax=”none” enable_mobile=”no” parallax_speed=”0.3″ video_mp4=”” video_webm=”” video_ogv=”” video_url=”” video_aspect_ratio=”16:9″ video_loop=”yes” video_mute=”yes” video_preview_image=”” border_size=”” border_color=”” border_style=”solid” margin_top=”” margin_bottom=”” padding_top=”” padding_right=”” padding_bottom=”” padding_left=””][fusion_builder_row][fusion_builder_column type=”1_1″ layout=”1_1″ background_position=”left top” background_color=”” border_size=”” border_color=”” border_style=”solid” border_position=”all” spacing=”yes” background_image=”” background_repeat=”no-repeat” padding_top=”” padding_right=”” padding_bottom=”” padding_left=”” margin_top=”0px” margin_bottom=”0px” class=”” id=”” animation_type=”” animation_speed=”0.3″ animation_direction=”left” hide_on_mobile=”small-visibility,medium-visibility,large-visibility” center_content=”no” last=”no” min_height=”” hover_type=”none” link=””][fusion_text columns=”” column_min_width=”” column_spacing=”” rule_style=”default” rule_size=”” rule_color=”” class=”” id=””]

You don’t know how much REAL and Usable Free Space you have.

The challenge with free space is not in identifying what is free or what is not. It is in the summarization and tracking where the information begins to lose its value, or where it creates a false sense of security and leads to poor decisions.   In order to make this real I will provide a “real world” example from a typical enterprise class customer.

The Environment:

An enterprise environment with about a dozen storage arrays from various manufacturers.  These arrays are located in 2 primary data centers – Production (PDC), and Secondary (SDC) and some arrays are located outside of all data centers in remote locations.

The standard unit of storage allocation for new storage being provisioned is 4 TB, so new storage requires this much space in a single pool or location in order for the volume to be provisioned.  The customer has some volumes “thick” provisioned and other volumes are using “thin” provisioning.  (Thick provisioning makes all allocated space as used and all free space will be invisible to the storage array)

The Problem:

The customer has 12 different storage arrays – since the customer is not using any storage virtualization software (like IBM Spectrum Virtualize) each array’s storage needs to be managed separately. (no storage sharing is supported across array boundaries.)

Each array may have more than one storage pools and these pools are typically managed separately as well, but inside the arrays storage pool sizes can typically (but now always) be adjusted dynamically depending on the specific array.  Each pool with an array therefore has it’s own free space, and then volume (or LUNs) of storage are allocated and assigned to specific a specific pool on an array.

An example of what this means in the real world:

12 storage arrays

6 pools/array average

10 volumes/pool

Pockets where free space = 12* 6 * 10 = 480 pockets of freespace within the enterprise

When thin provisioning is utilized consistently you can reduce the number of pockets of free space from 480 to 48 because all free space is then given back to the pool to be shared equally by all volumes in the pool.

The next set of challenges are created from the new “all flash” arrays on the market, like PURE Storage, which store everything in compressed and deduplication format.  These means that a 4 TB LUN may only require 1 TB of space on the array because all unnecessary data is not stored.  In this case the customer’s usable storage can be much higher than the raw storage provided by the array, and so a series of additional analytics needs to be performed to calculate the effective space the customer can use to store data and how much of that effective space can be considered free.

Since most enterprises are comprised of a mixture of “old and new” arrays having analytics and tools which adjust according to each array technology is critical.

Finally, we need to consider several things within the customer environment including – subtracting orphaned storage and the customer’s standard unit of storage allocation.  Orphaned storage is storage that allocated and being used but no hosts or NAS shares are mapped to this storage therefore is not in use.  The standard unit of allocation means to omit all storage blocks under this size because these storage block are not usable in the customer environment.

The Customer example:

The picture below is from an array with 23 TB of “freespace”.  However, upon a more detailed analysis the customer has 7 pools on the array and 5 of these pools have under 4 TB of freespace and therefore we would omit that storage from our available freespace.  Each of the pools with greater than 4 TB has room for 1 volume in each pool – this means the effective freespace within this array is actually 8 TB, not 23 TB – a dramatic difference that is not accounted for by just summing all freespace within a storage array.

How Visual One Reports and tracks ALL levels of free space across the enterprise:

Visual One Provides online and drill down capabilities at all levels of storage reporting including:

  • Enterprise view
  • Data center view
  • Array view
  • Pool view
  • Host & LUN View
  • Business Unit views

This allows customers to view and track storage usage without complex report generation, understand both: where they are and what their trends are, in order to be able to track as well as project clearly when (or if) additional storage is needed.

Best Practice Recommendations:  (to reduce freespace fragmentation)

  • Utilize Thin Provisioning wherever possible
    • Allows freespace to be managed at the pool level versus inside the volume level
  • Keep the # of pools on an array to a minimum (4 or less if possible)
    • The lower the # the more efficient space utilization can be achieved)
  • Utilize a storage virtualization software
    • Provides vendor independent storage implementation
    • Allows storage pools across multiple arrays – array boundaries are removed or minimized

[/fusion_text][/fusion_builder_column][fusion_builder_column type=”1_1″ layout=”1_1″ spacing=”” center_content=”no” link=”” target=”_self” min_height=”” hide_on_mobile=”small-visibility,medium-visibility,large-visibility” class=”” id=”” background_color=”” background_image=”” background_position=”left top” background_repeat=”no-repeat” hover_type=”none” border_size=”0″ border_color=”” border_style=”solid” border_position=”all” padding_top=”” padding_right=”” padding_bottom=”” padding_left=”” dimension_margin=”” animation_type=”” animation_direction=”left” animation_speed=”0.3″ animation_offset=”” last=”no”][fusion_separator style_type=”none” hide_on_mobile=”small-visibility,medium-visibility,large-visibility” class=”” id=”” sep_color=”” top_margin=”30″ bottom_margin=”” border_size=”0″ icon=”” icon_circle=”” icon_circle_color=”” width=”” alignment=”center” /][/fusion_builder_column][/fusion_builder_row][/fusion_builder_container]