localdata

Everyone will look for all-in-one but end up with a portfolio

no single method without a single data store. Our protection methods are as scattered as our data

Protection has to be more policy-based - call it imperative vs. declarative. Have systems smart enough that you can tell them the desired end state in business relevant terms and they ensure policy compliance.

fit for purpose has taken over, platforms have become ez enough to manage, no longer necessary to make a platform do something it's not truly built for because that's what you've already invested or have a skill set in

To be truly holistic, it's likely a portfolio approach simply due to all the different levels in the stack right now. Some companies have grand aspirations there (mine included) but there's so many potential data layers to handle.

Evolving from infrastructure centric (LUN/VM) to app-centric. That's how you eliminate the silos.

The challenge becomes keeping operational simplicity as expand to solve more and more business problems (aka handle more data sources). That's a HARD challenge.

The world of motherhood and apple pie - a single solution that does everything. We could always go back to mainframes.

@csislive Wait isn't the cloud a software mainframe :-)

for some reason, it feels like I should say "single pane of glass" right now. :)

i see a transition from legacy solutions and architectures to cloud friendly solutions and architectures for protecting data

@andriven, that's great, had same feeling

The real innovation will be a portfolio where we automate as much as possible and avoid human interaction.

that address scalability, costs in the cloud,

World changing so fast, portfolio of BoB seems today's normal. Definitely higher propensity from customers to deploy BoB as the expense of single throat to choke.

i sometimes think of the single pane of glass as a meta of a meta

I think there is a mainframe migration to AWS - double rainbow

@dinisco @andriven "single pain of glass"

@csislive the scary innovation is when the machines take over ;-)

@csislive, yup, it's never worked in the history of IT

@csislive skynet is coming ;-)

@chris_dag - How's the yurt - do you have a rack of gear in there yet?

@michaelcolbypro It's a bet on where it starts Facebook or Google.

i think that certain categories of data engines can be protected by the same solutions

@PeterSmails App-centric still creates silos. The person who owns SAP doesn't always talk to the guy who owns proprietary apps. Infrastructure can provide a common ground. Agree its not LUNS/VM, though!

@csislive sadly i agree. one day my echo is going to talk back to me. Sirius is going to tell me off

@csislive i mean Siri.

the biggest challenge to data movement still is the speed of light

Security, Accessibility, and Operational Overhead.

cost, bring the data to burstable compute is a great concept, works really well with known and manageable data sets, very tough when working sets are hard to identify

if you have compliance requirements, tracking where your data lives can be very time-consuming and costly if not done well.

security for sure. manageability. containing costs

Data Protection, Adhering to compliance

Good luck with GDPR with data scatted in so many places.

provenance (version control) and secured access plus just the 'weight' of moving all that data ...

Clouds are the new silo. Data mobility is a huge challenge. Workloads dictate data placement (e.g. test/dev, performance, BI) but only if you can intelligently move data sets.

under the heading of be careful of what you wish for, you may get it. Data is more accessible, can be generated far more quckly. but that creates new challenges

Knowing what data you have where and understanding how to protect it. Oh and moving it around. WAN networks are still behind in our ability to generate data. Even cloud is a challenge to make sure you have data in multiple regions.

I see 4 data management challenges: data protection, data movement, search & discovery, and learning - the use of the data to extract value.

we start by analyzing requirements and selecting the best platform when architecting for data at scale, now we run cost models before we do anything

absolutely. And yet we have to explain to people, still, the importance and need for backup

@storageswiss Ha! GDPR definitely flies in the face of current data dispersion patterns.

its the old "replication is not backup" discussion all over again.

@PeterSmails Totally agree with the data movement point. Got to move data from where it is to where it needs to be. Cloud bursting is in its nascency but under a lot of discussion.

DM pillars include Protection as table stakes (no matter where), mobility (gotta break the silos), monetization (leverage the data in native format anywhere).

In the future we have to not care where data is located, just know that it's discoverable, protected, and describable (is that a word?).

@andriven definitely agree with the compliance implications. Seems that for certain apps, most of the architecture discussions center around a few key compliance topics.

ironically the biggest cause of the need for backup and data protection is "human error"

systems and data need to be protected from the people who use and manage it ;-)

@MichaelColbyPro Agree that hybrid and multi-clouds totally change the backup discussion.

@MichaelColbyPro Agree. As you replicate, you replicate over your database corruptions, viruses, and accidental deletions! Ouch.

@steve_pao So much yes to that - especially when regulations by definition move slower than technology (can't be avoided). Riddle me this - when will specific language requiring tape be removed from some EMEA regulations?

@csislive Yes. Describable is a word. http://www.dictionar...

@andriven Funny how tape got written into regulation in the storage space; just like how packet inspection gotten written in on the networking side!