StoragePredictions

1,2,3 stop throwing FUD around, it's not pretty

Focus on value, value and more value...

- 1)backup goes away and traditional storage is more reliable and protection is a part of normal data services - all in 1

continue to not lose stuff, be more composable [sw], be more network aware

I wish for: 1) WORM flash that is dirt cheap. 2) Encryption made easy (and so that I and only I can ready my data) for cloud storage 3) NFS 4.1 everywhere.

2)storage becomes invisible

Buck up everyone, there is pretty awesome things still left to be done in storage. Deep analytics for one!

3) FLAPE baby - flash and tape

@virtualirfan I'd love to see DB logs go away ala ZFS with append only styles!

I think 1, Simplicity (gone are the days of complex storage architectures. 2, Agility being able to move data between different cloud Private to Public etc, 3, FUD Removal

@skenniston Didn't strongbox do that?

Drive level "user" upgradeable deduplication, compression, encryption firmwares?

@skenniston And we need one more layer of abstraction to achieve that!

make flash cheap enough that you don't need deduplication... wait 3D NAND already did that.

@virtualirfan Who cares about analytics in storage when you're struggling just to buy enough storage to meet requirements? Analytics are for the rich and the huge, not the 95% of regular businesses. Storage digital divide there.

@storageswiss Dedup never goes away since it improves margins for the vendor

how about native data API

@cakeis_not_alie No, you are wrong. Every smart maturing industry looks to analytics for growth. We need to absorb that!

@cakeis_not_alie 1. A universal agreement that pNFS was a bad idea. 2. The overhyped death of FCoE 3. A truce in the arms race of unrealistic dedupe and data reduction vendor claims.

sorry... how about standard HTTPS/API set that could manage information puts/gets... S3 is cool, but as we move from CRUD to CreateReadAppendQuery, not quite complete....

@furrier With NTAP and Nimble suffering startups on SAN storage File storage will NOT do well. Also, Object storage and VMware optimized storage companies will have difficulty.

Can I have everything in the Venn diagram? 1:Fast 2:Highly available 3:Cheap :)

@virtualirfan vendors need to stop worrying about margins and focus on user needs - they'll be better off in the long run

NFS is really broken as you lose attribution!

that storage becomes just as invisible as compute.

@alexgalbraith :) good one!

@smuddu we've seen a bit of a run on object with Amplidata and Cleversafe getting taken out...

meta data and meta data management is the next thing

@virtualirfan Vendors looking to find a new path to high margins look to analytics for growth. Ordinary companies look to analytics to analyse how best to take advantage of THEIR customers, not be taken advantage of by their vendors.

@skenniston Agreed. And users are suffering with storage systems today because of lack of intelligence, automation, troubleshooting, all helped if we wake up and do more analytics, automatic remediation, etc.

1. Quality of Service (end-to-end), 2. Analytics and 3. Flape @skenniston

@smuddu how are you! you're right objects make a ton of sense!

@virtualirfan "Analytics" needs to deliver real value, not just make pretty graphs that don't bring more dollars in the door. They certainly need to not be taking more dollars OUT the door.

@storageswiss what ever happened with WORM laser cubes.. tape still SUCKS

@storageswiss Can we standardize on calling 7.2K drives low latency tape?

@cakeis_not_alie Ordinary companies are just vendors for other companies dude.

@cakeis_not_alie Agreed with you there! Right now they tell you that the patient is bleeding out, they need to tell you how to fix

@cakeis_not_alie Pretty shiny graphs bring more dollars (or real money) to those that sell "Analytics" ;-)

Need Azure, Oracle, IBM, Google to offer as good as AWS S3 compatible storage Service.

@skenniston Amen. I need: 1) to store my bits 2) to get access to my bits fast 3) to never lose my bits, even if a disaster happens 4) from 4) to do all of the above AS CHEAPLY AS POSSIBLE. That's it. That's what i need from storage. Do that. Do it well

@Lost_Signal No. Because you can't automatically eject them and have them sit on a shelf for 30 years

@virtualirfan Yes, and? I don't want my vendors charging me a cent more than is absolutely necessary. Meanwhile, I want all the money possible from my customers. That's business. So I squeeze my vendors and optimize my clients.

I'm curious what % of 10 year old tape ever gets read.

@furrier MAPR need to decide if they want to be storage company of HDFS or Hadoop company. WIth Hortonworks market cap at $700M MAPR will be under lot of pressure to make decisions..

@Lost_Signal in some industries - a lot

@virtualirfan As a storage CSUTOMER, what does storage analytics do for me? What's the value for the dollar? What's the advantage to me over cheaper storage that's faster with more capacity?

@furrier DO enterprise customers need another Backup storage company beyond Data Domain? Not sure there is opportunity to for storage backup for Hadoop, Cassendra ?

@Lost_Signal or how much 10yr old tape CAN still be read!

@cakeis_not_alie Whoever in storage figures out that helping customers buy only what they need, when they need it with a product that fixes problems before they get in the way will win. That's analytics end to end.

@cakeis_not_alie On large, structured storage infrastructure it allows you to break that infrastructure in smaller, cheaper and more efficient storage units. Seen that recently. Don't need 300Tb storage, need 10 x 30Tb storage.

collaboration collaboration collaboration 1) standards around storage API's 2) better encryption 3) enhanced application snapshot support to be extended deep into the application stack.

@virtualirfan No vendor wants to "help a customer buy only what they need". Are you nuts? Vendors want to find out how to convince a customer to buy as much as possible. Exact opposite of "only what they need".

@cakeis_not_alie Your premise is wrong. There is no such thing as cheaper storage that's also faster with also has more capacity. It doesn't exit.

@virtualirfan Storage data or content analytics is not going to happen in next several years. Analytics may be used for monitoring/system level diagnostics mainly

@virtualirfan Now, a customer might employ analytics to ensure they only buy what they need, but only if the analytics don't cost more than they save.

@cakeis_not_alie Duh!

@virtualirfan agree.. and the storage vendors need to orient around the natural value decay factors of data... how long is it really valuable, when is it most valuable.... unfortunately the stream->store pipelines are not fully exploited by most

@virtualirfan Funny, I make a living building cheaper storage that's faster and has more capacity. For that matter, talk to HDS about how they're doing selling against EMC. It DOES exist. That's what the storage wars have been about.

@smuddu Agree that's conventional wisdom. However, we can at least figure out that as the right path ... some startup will execute to it

@cakeis_not_alie Trevor can build storage that is faster&bigger data but Trevor is expensive :-) And vendors will compete against each other on cost / TCO all day long.

@Vecchi_Paolo checkout https://github.com/G... new open source graph framework!