Loading...

Employing Cloud Blob Storage Smartly To Match Your Data Copy Needs

 

Blob Cloud Storage Service

HOT Tier

Blob Cloud Storage Service

COOL Tier

Blob Cloud Storage Service

COLD Tier

Overview

Data that is expected to be frequently accessed.

This is best suited to data copies that have very short retention or will be actively read on a repeated basis.

  1. Primary Backup copy roles, as that is normally a data set that will be read in (DASH or AUX Copy) or Dash-Full process.
  2. As a Secondary Data copy source short term retention (<30days) or for data sets that are routinely accessed. For example using LiveSync, Data Verification, or Activate/Indexing operations.

Data with longer retention needs and more limited access expectations.

We recommend aligning this with +30-day backup and DR copies. This service offers storage usage at 50% the storage cost of the Hot tier, but the user will pay an additional $/GB usage fee on each GB read.

Any data set that has more retention and minimal data read/access expectations for DR, discovery, individual restore, etc. is "generally" better suited to COOL if you plan on managing something more sizeable 10TB+. 

Most protection and archive users store much more data than they access monthly, which makes this a better set of economics for the use-case.

Azure assesses the minimum stay at 30days, so if you only apply a 7-day copy retention rule, we will delete those blobs as required but the user will still be assessed a minimum 30 day stay/storage usage charge.

Data with very long retention needs and the lowest frequency of access. This tier offers the cheapest storage usage fees, but the access fee can add up quickly and require several hours of delay as the data is recalled to the Hot/Cool access tier. This was designed to replace more traditional tape, COLD data storage use-cases.

We recommend using this for the long-term, last-data copy in your Storage Policy use-cases.  It should require a minimum of at least 6 months (180+ days) of retention and a very low probability for access.   If you deleted it or moved it to a higher tier prior to the 180-day period, you will still be charged for the minimum 180 day stay per Azure.

Accessing and recalling large amounts of data can be expensive $10K+, this incurs a higher GET fee $5/10K requests plus the Access fee. On average we estimate that to be $150/TB*.

Many customers have learned this mistake the hard way. When they prematurely froze the backup data into the COLD tier, and then they triggered large copy/read operations that drove the thaw/recall action and were faced with a high cost for the recall event.

Note: When recovering deduplicated data from archive/cold cloud storage, all the volumes associated with the Deduplication Database (DDB) are recalled. Therefore it is recommended that the DDB must be sealed periodically (for example every six months) to reduce the amount of data that is recalled.

Planning Your Usage Charges

  • Storage Usage Cost: Yes,~$0.02/GB/Month
  • Read Access Cost: None
  • Zone Replication: Yes, add-on cost to replicate the blob store
  • Storage Usage Cost: Yes,~$0.01/GB/Month
  • Read Access Cost: Yes, $0.01/GB, $10/TB
  • Zone Replication: Yes, add-on cost to replicate the blob store
  • Storage Usage Cost: Yes,~$0.001/GB/Month
  • Read Access Cost: Yes, Access+Read API cost= *$150/TB

Best Fit For

  • Data that you expect to actively process with indexing/analytics or short backup copy needs. Best suited for Primary Backup copy
  • Short-term data to be retained for under 30days
  • Copy with higher probability of retrieval, data is readily available in the storage bucket
  • Backup Retention Copies or Active data for DR or other needs. Best suited as the Secondary/Offsite Backup copy
  • Data to be retained for 30 days or more
  • Probability of retrieval is modest; data is readily available for reads
  • Deep compliance retention data (after the active discovery period), archive data that is older than 180-days or record keeping data are all strong candidates. This is where many plan to shift older Tape Vaulting datasets into the Cloud COLD archives.
  • Probability of retrieval is lowest, and you need to plan for up to 15 hours for recalls occurring, before you can access the data

Last modified: 12/9/2019 9:33:03 PM