IBM Spectrum Scale provides a unified storage platform for multi-protocol access like SMB, NFS, posix, Swift, S3 API and also HDFS. With this unified platform a variety of interesting use cases can be addressed. This blog is an attempt to let you know what you can do with your Spectrum Scale unified object storage. Here are 10 cool things I feel can be done-

1. Thin-thick storage capacity site deployments for object data

Spectrum Scale object when configured with Active File Management (AFM) gives the capability of caching the object data across remote sites with thin/thick storage capacity. This give an active-passive behavior where the active site is of lower capacity than the passive/backup site. AFM provides a persistent cache that also works in disconnected mode thus providing uninterrupted IO. This has also been successfully deployed by a customer, check the report at – https://www.linkedin.com/pulse/ibm-spectrum-scale-makes-news-japan-smita-raut?trk=prof-post

2. In-place analytics over object data

Most of the popular analytics software work over a filesystem and to analyze object data, typically data needs to be moved from object store to a filesystem, or gateways have to be provided over object stores for file access. With Spectrum Scale’s unified file and object access feature, same data can be accessed as object as well as file thus eliminating the need for data movement for analytics. Check out a presentation on it here- http://www.slideshare.net/SmitaRaut/spectrum-scale-object-analytics

3. Automated tiering of hot objects to faster data pools

You would always want your most frequently accessed or hot objects to be placed on fastest storage pool like SSDs and the cold ones on slower pools. Spectrum Scale provides automated heat based tiering that can be configured for objects. For more details refer- https://smitaraut.wordpress.com/2016/06/16/hot-cakes-or-hot-objects-they-better-be-served-fast/?cm_mc_uid=50742467362414556086843&cm_mc_sid_50200000=1486447793

4. Multi-Region deployment with active-active configuration

When client access to object data is over a WAN or a high-latency network, the network can introduce an unacceptable delay and affect quality-of-service metrics. To improve that response time, you can create a replica of the data in a cluster closer to the clients using the active-active multi-region replication. For more details refer- https://www.ibm.com/support/knowledgecenter/STXKQY_4.2.0/com.ibm.spectrum.scale.v4r2.ins.doc/bl1ins_multiregionoverview.htm

5. Writing your own pre/post object processing software (middleware)

Spectrum Scale Object is based on Openstack Swift implementation and you can write your own middleware in proxy/container/object pipeline for doing pre/post processing of objects. One of the examples that IBM has open sourced is a middleware that analyses object data using IBM Watson cognitive services to derive insights about it and automatically add meaningful metadata to the object. More details on the opensource project, how it is done and a demo here- http://www.slideshare.net/SmitaRaut/spectrum-scalecongnitive. But before writing your own middleware and putting it on Spectrum Scale cluster, please contact IBM.

6. Moving your object data to tape

Many financial organizations like banks in India mandate tape out even if the data is object. Spectrum Scale supports tape integration that works for object data. For more details check the redpaper on this topic- http://www.redbooks.ibm.com/redpapers/pdfs/redp5237.pdf

7. Enabling object access on your existing files

With 4.2.2 release Spectrum Scale also provides a capability of enabling object access to your old existing file data. See how to do that here- https://www.ibm.com/support/knowledgecenter/STXKQY_4.2.2/com.ibm.spectrum.scale.v4r22.doc/bl1adm_enablingobjectaccessonexistingfilesets.htm

8. Tiering based on object metadata

You may have tagged data that is most valuable to you by means of some metadata and would want this to reside on faster storage pools. By means of ILM capability of Spectrum Scale it is possible to tier your object data based on its metadata. Check how this can be done here- https://developer.ibm.com/storage/2016/09/08/are-you-heating-up-your-frozen-daiquiri-and-even-pay-for-it-serve-the-as-you-specified-hot-objects-fast-and-move-the-ones-you-want-to-freeze-into-the-fridge/

9. Encrypting your objects on disk for better security

Objects in Spectrum Scale can also be encrypted. You need to create an encryption policy for that- https://www.ibm.com/support/knowledgecenter/STXKQY_4.2.2/com.ibm.spectrum.scale.v4r22.doc/bl1adm_storagepolicyencrypt.htm

10. Compressing your objects on disk for storage space optimization

Objects in Spectrum Scale can also be compressed. You need to create a compression policy for that- https://www.ibm.com/support/knowledgecenter/STXKQY_4.2.2/com.ibm.spectrum.scale.v4r22.doc/bl1adm_storagepolicycomp.htm

(Views presented here are my own and not of my employer’s)

2 comments on"10 cool things you can do with IBM Spectrum Scale Unified File and Object storage"

  1. […] 10 cool things you can do with IBM Spectrum Scale Unified File and Object storage […]

  2. […] Explore 10 cool things you can do with IBM Spectrum Scale Unified File and Object storage […]

Join The Discussion

Your email address will not be published. Required fields are marked *