HybridStorage

If the answer is "I'm depending on my vendors for this" then the question very quickly needs to be "Are you sure you're talking to the right vendors?"

- Step 0 is for enterprises to become more self-aware about how poor their baseline data is. "Our CMDB is perfect". Ok. Sure it is. Have to honestly assess what you have before you can predict what you do next.

1. Business Case 2. Data profiling 3. Applications Rationalization. All this with some steering from experienced hands/brains.

@ballen_clt - A few companies are finally starting to realize that they need to assume that most data is wrong or atleast stale. So they need to focus on discovery and dependency mapping. Another step is to look at apps not just raw resources. The biz doesn't care about servers..

@ballen_clt Can't act on what you don't know! Worse, if what you know is wrong, hard to make the action right.

great question - the industry didn't do great at managing a multi-vendor environment, will it provide better tools in a #hybridcloud world? It's too easy to allow things to get complicated fast. Let's hope we've learned from the past.

@sarbjeetjohal This goes to my point about "vendor lock-in". The answer has to be best for the business while also allowing me to move between technologies as needed without a massive overhaul of my infrastructure. Businesses today can't afford to be locked in to any platforms.

@stu Used to be involved in SNIA and watching vendors drive decisions so they could hold on to their customers and NOT support multi-vendor was scary. SMI-S struggled because of vendor politics from a NE-based vendor.

Here's a checklist to assess your cloud strategy and see if there are ways to improve IT operations in your enterprise: https://www.hpe.com/us/en/pdfViewer.html?docId=a00...

Customers often talk about an automated support experience thanks to global AIOps to improve IT operations... less time on the phone, more time innovating and supporting new-age applications!

Companies that I talk to are using machine learning and that requires working knowledge of the data and data flows. Impacts are wide reaching from security to user experience

The ability to predict and proactively prevent disruptions thru AI/ML is becoming a game changer for customers. Of course, that's just the starting point. The value extends into keeping IT infrastructure optimized.

@channelmkt1 Nice!

HPE InfoSight gets IT out of the hot seat. Instead of fighting fires and dealing with disruptions, this AI platform looks across the entire stack, leverages AI/ML to predict/prevent issues

@storagesandeep That sounds like intelligence!

use of #AI and #ML will be plugged in from the way you write code, to way you deploy it, way you mine logs in operations to way data gets into its own storage tiers. Can’t pinpoint single tools, yet 😉. Complex issues are a big set of small ones (in IT at least)

And delivers prescriptive insights for Storage and VM Admins. It gives you the super power to find the needle in a haystack.

Check out this video - sleep better at night with AI

https://www.hpe.com/us/en/solutions/infosight.brig... https://www.crowdchat.net/s/55v4r

(edited)

And love that we've extended InfoSight to 3PAR, HPE Primera, HPE Servers (ProLiant, Apollo, Synergy compute), and now HPE SimpliVity!

@furrier yep, ML has huge implications on security and then storage... compute is already well handled (still can benefit from ML)

Unless you're in the IT business AI & automation should only make your business better & more efficient. Technology today should improve the businesses time-to-market while lowering the overall costs. And "vendor lock-in" has to be a thing of the past here as well.

@sarbjeetjohal ML and servers is developing but taking that insight and feeding it into an AI engine like InfoSight, not aware of anyone else doing that today.

This is happening as we speak. Anything that isn't designed to be "cloud-oriented" today is basically a brick.

1. Policy based storage orchestration

The real difference is who understands the end-to-end nature of the application & how best to optimize it holistically. That becomes a challenge for single-component vendors.

Global Intelligence via AI-driven infrastructure operations is key for optimizing cloud oriented resources

AIOPs Meetup next month in Boston! https://www.meetup.com/VirtGBoston/events/26417773...

@storagesandeep "Global Intelligence" = BINGO. Need to understand the life of the data in order to ensure the most optimal delivery & access methods. Be it public, private, on-prem etc infrastructure.

2. ML enablement on telemetry data 3. Applications rationalization. BTW it all

bi-directional data mobility, automated policies, unified data services across public cloud and on-premise with AI that drives decision making

it all starts with data profiling...

I don't particularly like sharing content that you have to sign up to get but this "Multicloud Storage for Dummies" is really good. https://www.hpe.com/us/en/resources/storage/multic...

- Analytics and automation will help address over-provisioning and waste. The next step will hopefully be for tech to also predict the right type and tier of storage for data. (You're using block but object would be better. Put this data on flash vs. NVME)

@davidwang I'd like to vote for your comment twice!

By gaining global knowledge of the underlying application characteristics and it's interaction with the data architecture. The key will be the ability to make real-time changes based on where the data sits, how it's accessed and how it's protected

A distiributed graph of storage resources will pervade the edge-to-edge multicloud. AIops and policies will automate the continuous edge-to-edge hub-to-hub cloud-to-cloud persistence of data across this graph.

@jameskobielus Great insight. Maybe include "a distributed graph of data resources," in keeping with some of the observations. Both will be required: a network of data --> advanced data services.

@ballen_clt I like where you're going.... In theory, advances in technology and management should leverage AI to determine not just where your data belongs, but also the specific type of storage that is best suited for your data

@jameskobielus well said!

yes sir.

automation, machine intelligence, analytics

Single pane of glass management across on-prem and public cloud environment.

The real question, how is "ease-of-use" defined? Is it defined at the the infrastructure layer, the individual component layer? By customers? End Users? Depending on where you're standing you might have a different answer.

if you start from the app point of view, you need consistency of reliability and data services for existing mission critical apps to leverage cloud storage and ultimately in a hybrid cloud fashion

Developing a consistent data architecture across hybrid-cloud is painful but awesome hygiene for long-term ease of use. It's tempting to go all in on an on-prem solution or an all public cloud solution. Great short term benefit but long-term it hurts all aspects of ops.

Here's a short and sweet guide our team put together that is a "health check" for cloud ready storage. Definitely worth checking out! https://www.hpe.com/us/en/pdfViewer.html?docId=a00...

@dvellante Those are great buzz words but how do they equate to "ease-of-use" really? Many vendors claim to check all those boxes & yet fall flat on the actual "easy" part.

make infrastructure manage itself, if cars can drive themselves, then why not hybrid cloud, right? i don't want to spend time sitting in traffic, and i dont' want to spend the time tuning my storage #autonomousbaby

single control plane, ML capable, policy driven orchestration and most presentable telemetry reports/dashboards which give us visibility in price/performance ratios. Yea, that’s a dream list.

Great insight Jorge! As more folks start to draw stronger lines between digital capabilities, data, storage, and cloud, it means ease-of-use concerns expand to more use cases, don't you think?

The fact is an "easy storage array" is just that. It doesn't necessarily make my application or the infrastructure itself simpler. True "ease-of-use" needs to be resolved at the top layer with all the components underneath doing their job efficiently & invisibly.

And let's not forget the Support Experience. The "Set-it-and-forget-it" promise of the public cloud experience is driving adoption. #HPEInfosight is really the key here.. not just as an AIOps, predictive analytics platform but also as a support automation engine

(edited)

@sarbjeetjohal - You nailed it. Ease of use has to hit atleast two audiences. Single pane of glass for both operators and decisions makers. "Hands on the keyboard" have different requirements than executives who need to defend strategy and look towards the future.

Jorge these 3 items directly attack the IT labor problem. Automation reduces provisioning, troubleshooting & tuning costs, machine intelligence allows predictions & action (in some cases) & analytics provide a basis for continuous improvement. AWS has implemented these buzzwords

@dvellante Completely agree, that was actually my point. Those buzz words without "IT labor problem" attached are just buzz words. To make it real, they MUST address IT labor in a positive way.

We can't take for granted the all vendors are addressing IT labor in their products. Many of them aren't. They're using the buzz words but either solving only one very specific piece of the problem or missing IT labor entirely.