#StorageSummit - CrowdChat

Keeping Data Forever

Wikibon & Oracle discuss the issues around keeping data forever, incl. tape v. cloud archive debate

#storagesummit @dvellante @furrier @oraclehardware oracle backup storage

Register for Storage Summit Here

LeaderBoard

#StorageSummitStorage BackupWikibon and storage experts discuss how to meet backup SLAs and reduce costs.

#StorageSummitVirtualization & Storage Storage experts discuss how server virtualization affects storage.

Stream Ended

Establishing a secure connection

Oracle Hardware21

In considering deep archiving in the cloud, Amazon Glacier for example, vs. onsite tape what cost/benefit tradeoffs should you look at?

[+] Show Hidden Comments

There is no ROI with large scale cloud storage with out Tape..

Oracle Hardware

The Rise of the Machines: Rebirth of Tape: http://wikibon.org/wiki/v/Rise_of_the_Machines:_The_Rebirth_of_Tape

Storage Alchemist

Amen to that Robert!

In fact well-known internet companies use tape as the ultimate backup.

this tape is dead is ridiculous, tape is cheap. A $70-75 list price for LTO-6 works out at about three cents a gigabyte. Consumer-grade 2TB SATA drives are only now starting to approach that sort of price

@cmosoares the argument most people make is SSD is the way to go but still from the data is more expensive than tape by a mile

Oracle Hardware

Google reveals TAPE-TASTIC data centre in saucy vid: http://www.theregister.co.uk/2013/01/28/google_oracle/

I was working with a customer to migrate a DB.. they didn't have enough disk space to make an extra copy.. (it was a big DB).. I convinced them to use our Tape. We wrote 660MB a sec, then read it back @ 735 (only 3 10kc drives..) amazing..

Oracle Hardware

Do you include factors such as cloud ingest, storage and retrieval (could take several hours), network bandwidth or mail tape/disks vs onsite disk/tape libraries, media, energy, management resources, footprint.

If you're looking at Glacier, you're not really in a hurry to get your data back (5-6 hour SLA). Hopefully that's not your backup strategy. Probably a good differentiator between backup and archiving.

Oracle Hardware19

Studies show that long-term storage environments using just tape or mixture of disk and tape have lower TCO than all-disk. What’s your experience?

[+] Show Hidden Comments

Periodic backup technologies are well suited for data retention and compliance, but many have issues relying on them for business continuity - HUGE issue to address

Oracle Hardware

Infographic: http://www.oracle-downloads.com/StorageTekT10000DInfographic/

Oracle Hardware

DCIG rates Oracle StorageTek SL8500 Tape Library #1 http://www.dcig.com/2014/03/dcig-2014-15-big-data-tape-library-buyers-guide-now-available.html

DCIG 2014-15 Big Data Tape Library Buyer's Guide Now Available - DCIG
DCIG is pleased to announce the availability of its DCIG 2014-15 Big Data Tape Library Buyer's Guide that weights, scores and ranks over 80 features on more than 40 tape libraries from eight (8) different storage providers. Driven by growing corporat...

tape archive solutions are proven to stand the test of time for long term data retention

Storage Alchemist

@CrowdCaptain - one note, i know we use the word "archiving" to mean many things - i do think its important to note that now w/ new tape technologies such as LTFS, you can use tape for scale-out NAS, cloud, long term retention - many more use cases

Need to be careful to differentiate between backup and archiving.

Storage Alchemist

IT needs a balance, Flash/SSD for performance (best $/IOP) and tape for long term retention (best $/GB) - Long live Flape - it will be the middleware tools - getting the data to and fro along w/ metadata mgt tools that drive this success

Oracle Hardware

LTFS for Dummies: https://go.oracle.com/LP=982/?elqCampaignId=1776

I expect more activity around data tiering in the cloud with cloud tiers ranging from enterprise-class to archive

That is true -- but it does not require a "Long-Term" The Cheese moved in the Tape world after Oracle got a hold of the StorageTek Assets (via Sun).

“..not only can we can get it back in a timely and verifiable way, but we can get any version of it back for the user...like a giant time machine,” http://www.oracle.com/us/corporate/customers/customersearch/qld-brain-institute-1-storagetek-cs-1981825.html

Customer and Partner Search Queensland Brain Institute Increases Archive Capacity, Supports Limitless Innovation and Increases Data Response Speeds by...
Queensland Brain Institute uses Oracle’s StorageTek SL3000 modular library system and StorageTek T10000C tape drive for an optimized, tiered storage infrastructure to increase storage capacity and data integrity while minimizing cost and data center ...

“..not only can we can get it back in a timely and verifiable way, but we can get any version of it back for the user...like a giant time machine,” http://www.oracle.com/us/corporate/customers/customersearch/qld-brain-institute-1-storagetek-cs-1981825.html

Customer and Partner Search Queensland Brain Institute Increases Archive Capacity, Supports Limitless Innovation and Increases Data Response Speeds by...
Queensland Brain Institute uses Oracle’s StorageTek SL3000 modular library system and StorageTek T10000C tape drive for an optimized, tiered storage infrastructure to increase storage capacity and data integrity while minimizing cost and data center ...

Oracle Hardware19

Looks like things are settled again. Next question: Do you keep long-term archival data onsite? On disk, on tape? Offsite? On a deep cloud archive, such as Amazon Glacier?

[+] Show Hidden Comments

Storage Alchemist

good question - lets 'one up' - how do you know what data to put where?

What I really find odd - is I rarely find customers who balance the MTBF times against how long they want to keep data on a given media..

is archiving still just an insurance for recovery in catastrophic failure or are more customers finding ways to get value from the "data lake" through analytics, etc?

Oracle Hardware

Here are Wikibon's pros and cons for using Amazon: http://wikibon.org/wiki/v/Cloud_Computing_2013:_The_Amazon_Gorilla_Invades_the_Enterprise

Storage Alchemist

@oraclehardware also, data movement is the most expensive thing in the enterprise, if that is the case, what is the most economical way to move data around?

I find many want online vs tape where it's a "green field" and tape for where they have existing processes

Storage Alchemist

@CrowdCaptain today w/ LTFS you can have data "on tape" AND "on-line"

Storage Alchemist

@stu archiving is archiving - backup (to a secondary device or medium) is used for catastrophic failure

I like that report already posted below - it really raises the fact this is a multi-million dollar question with tape worthy of consideration. http://www.oracle.com/us/corporate/analystreports/industries/clipper-tco-storage-2013-1959019.pdf

A tiered architecture (D2D2T) provides lots of flexibility for where to put data

Oracle Database 12c with ADO+HCC gives you policy-based data placement and different compression levels for hot vs. aged data.

Oracle Hardware15

A recent report showed that the proportion of unstructured data compared to structured or transactional data is increasing in the enterprise. Are you archiving more images, videos, emails, sensor data, PDFs, etc.?

[+] Show Hidden Comments

Oracle Hardware

BMW turning to the cloud as it collects sensor data (M2M) from every car it makes – 10 million cars by 2017 will collect 1PB a day. http://www.gen-i.co.nz/insights/blog/dealing-with-petabytes-of-mobile-data/

Dealing with petabytes of mobile data » Gen-i New Zealand
You do not have JavaScript enabled. The Gen-i website requires JavaScript to load content and will not function correctly. You can find out more about JavaScript here.

I hear this all the time 'back up everything bc we might need it later' - so many cios that I talk to say 'back it all up'

As the internet of things heats up, companies will be drowning in sensor data.

Oracle Hardware

Infographic from @Wikibon shows data growth and sources of data http://wikibon.org/blog/data-footprints/

Data Footprints by Generations [Infographic]
Posted by Stuart Miniman in Big Data, Flash, Wikibon on January 10, 2013

Oracle Hardware

Infographic from @Wikibon shows rapid growth of unstructured data http://wikibon.org/blog/unstructured-data/

Revisited: The Rapid Growth in Unstructured Data
Posted by David Vellante in Cloud Computing, Storage, Wikibon on August 17, 2010

@cmosoares internet of things adds more urgency to the importance of all pieces of data bc it can be measured to provide a piece of insight & value for business improvement

@furrier Certainly adds new meaning to big data. Billions (trillions) of sensors delivering a continuous stream of data. Requires big machines to analyze and repositories for historical analysis.

Media & entertainment is another great example of burgeoning unstructured data growth. They used to keep a couple gold copies around - now retaining all footage from all cameras for future releases.

@KerstinWoods at #NAB it was overwhelming that all media use tape as back up

Oracle Hardware14

So let's get things started. Data has a short shelf life. It’s rarely or never touched after 90 days. Is managing old data a big issue for your company? How do you deal with it?

[+] Show Hidden Comments

Most of the customers I find - are literally scared of their aged data.. And the steer clear of it.. (leaving it in place).

Oracle Hardware

What makes them scared of their aged data?

one issue is the policy of the data will be reused and what about security

With many advancements in data analysis, companies are considering the future untapped value in historical data and forever-retention is becoming a reality.

Storage Alchemist

if they move it, then need it again and can't find it because something "happened" they are in trouble

Do these customers just keep the old data on primary disk?

If they leave their aged data in place they are consuming expensive storage resources and can reduce their overall storage costs and RECLAIM disk space by migrating aged data to highly economical tape storage

From talking to customers, the fear sits between Legal worries - and them just not having an official company policy about how to handle aged data

Oracle Hardware

what technologies exist to help organizations keep track of their aged data if/when it "moves"?

Storage Alchemist

@KerstinWoods I buy the data analytic play all day long, @dvellante

@Clipper Group study showed that disk-only long-term storage cost 26X more than tape
http://www.oracle.com/us/corporate/analystreports/industries/clipper-tco-storage-2013-1959019.pdf

Storage Alchemist

@Clipper Group study showed that disk-only long-term storage cost 26X more than tape
http://www.oracle.com/us/corporate/analystreports/industries/clipper-tco-storage-2013-1959019.pdf

Storage Alchemist

@KerstinWoods I buy the data analytics play all day long, @dvellante what about companies that are not in any sort of compliant business - there are more of that that have long term data storage needs....

Clipper Group study showed that disk-only long-term storage cost 26X more than tape
http://www.oracle.com/us/corporate/analystreports/industries/clipper-tco-storage-2013-1959019.pdf

Storage Alchemist

I believe that even thought LTFS is relatively new to the space, it really helps customers leverage the use of tape in a much more economical way...

@oraclehardwareClipper Group study showed that disk-only long-term storage cost 26X more than tape
http://www.oracle.com/us/corporate/analystreports/industries/clipper-tco-storage-2013-1959019.pdf