ibminterconnect

try not to move data around to suit server choices... ETL costs real money

Know what work the database does well and what it doesn't. Get your workloads into the sweet spot.

Consider ALL the business lines and people in an entire company, who don't know who may want access to the data. Don't create to small a scope, if the data is available, it may be utilized in ways not yet conceived of until a future date.

Don't limit yourself to a single software or hardware platform if you can generate substantial value with a change -- such as introducing OpenPOWER.

lots of focus on no-sql DB's, but need to focus on the business requirement and not chase the next shiny object

If you are using a hybrid cloud model ensure you have the support for your critical systems

Recruit, train, and incentivize all personnel with a common data literacy, and that increasingly focuses on fostering a new generation of self-taught "citizen data scientists."

Oh, and on a more simple level, be sure to get good backups (had to add that). ;)

focus on security of the data...

Increasingly with new breed of analytics - Do we all become Data Scientists ?

Adopt a hybrid data architecture that deploys fit-to-purpose data platforms--Hadoop, streaming, RDBMS, columnar, graph, distributed file, etc.--for different use cases, but within a common end-to-end architecture.

@GuyShone you as a data scientist... now that I would pay to see that...

@StevenDickens3 Balancing availability, confidentiality, and integrity (the CIA triad) is critical.

Evolve toward a hybrid cloud architecture in which public and private clouds are used where each is best suited, but with common security, governance, and ILM spanning them.

@GuyShone We will all need to have basic literacy in the core of data science--the need to build and test statistical models against empirical data--but we won't all need to become data scientists. However, more biz analysts will. Career path

Is there a 'look' Steve. Will try it out in Vegas

@majorhayden availability of 'systems of record' is key

The enterprise architecture conversation needs to start at the C-level. How will the CDO, CIO, CTO, CMO, etc. effectively use cognitive computing apps to drive disruptive outcomes across functional business areas?

where it should start and where it does are very different things. It should start with the LoB driving requirements

Sense sometimes C suite disconnected. Stong articulations of real value & #storytelling help engage #CEO #CFO early @IBMInterConnect

The enterprise architecture conversation should be a grassroots movement, starting with the people closest to the technology. They own the technology and know it best.

@majorhayden yes and no... shouldn't it start with e biz need?

@jameskobielus Perhaps it starts at both ends and meets in the middle management layer -- where things often get bogged down. ;)

sorry business need (fat fingers)

Enterprise architecture in the cognitive era is all about tapping into the power of AI, deep learning, machine learning, data science, & big data assets. How will enterprise build centers of excellence in these areas?

I agree with @majorhayden the movement is typically driven by the tech owners

@MsAngWelch it is.. but should it be?

It sounds like the responses are coming from different perspectives based on different understandings of the context.

can't build the plumping design until someone decides what the house need to look like...

@majorhayden It's not about "owning the technology." It's about owning the business outcomes. An enterprise architecture discussion needs to provide a planning framework for driving outcomes from shifting tech investments.

The C-level executive needs to meet with the Enterprise Application teams regularly to understand how the business data flows today, then bring in the Infrastructure team to formulate where they want to go in the future!

Enterprise architecture, t-shirts, and flip-flops can co-exist as well -- no need to be overly formal.

Chad has it...

@stevendickens3 true, there needs to be cohesiveness from the c-level down to the datacenter managers - they should be on the same page driving together

agree tech owners with an aggressive LoB sponsor but C-Level must guide and define boundaries

@chadlingmann Yes. The C-level executive ensures that the various biz-tech teams don't become isolated silos, but rather ca pool their efforts in service of cross-functional outcomes/apps.

Very important to not only asses the infrastructure but the inhouse skills to move forward

Finding ways to empower tech owners to influence C suite better is increasingly important as buying decisions involve more people

@MsAngWelch I think in the context of your question, Data Center managers are not as involved. They manage more of the physical planning and capacity to ensure they can operate what the Enterprise Architecture teams craft/purchase

got to get the teams to think outside their 'swim lane' and form cross functional small agile teams to focus on tactical successes

Cognitive advancements can generate value from data that was previously considered a drag on the finances of the company. We need to change how we assess value (and expense).

@StevenDickens3 Yes! Come to the executive table with a cross-dept approach.

@StevenDickens3 Yes. The new generation of developers in the cognitive era work in multidisciplinary teams: data scientists, data engineers, coders, subject matter experts, UX, hardware, etc. Creatives just as important as techies.

need to work cross boundary...

The culture shift in digital transformation is toward self-service, collaborative, distributed, cross-disciplinary teams that all have a common grounding in data science, cloud data services, and real-world experimentation

Agreed. And client / customer first which means get good at telling the story of 'what to do differently on Monday'. Transition plan rather than static blueprint

The culture shift in this new cognitive era is toward recognizing that the power of AI is in augmenting, accelerating, and extending developers' own natural abilities. The technology is a productivity tool.

mainframe is part of the discussion for sure, Cloud is not about a box architecture more about a consumption paradigm...

Mainframes, just like any other technology, is the sweet spot for certain workloads. Identify the workloads and get them into the right place.

lots of CIO's looking to make mainframe part of their cloud first strategy, be that either on or off premise, single or multi-tenant...

Who is to say mainframes don't run in the cloud? I would look at containers for portability, but I would not rule out mainframe.

@majorhayden 100% agree, lots of spurious chatter about re-platforming to go cloud, and that just adds risk. Just change the delivery model not the app

Mainframes are often the core repository for system of record enterprise data. In a cloud-first/data-first strategy, ask yourself how the mainframe's assets can be placed front and center while maintaining security, control, etc

If it's done right, still speed and reliability. http://thenewstack.i...

Think this goes back to clear story telling into Csuite to avoid assumptions of mutual exclusivity

Mainframe is a beast of a box for certain workloads be they systems of record or just apps that need 11/10 for security, performance or availability - with new offering such as mainframe-aaS in the cloud, clients have choices

Mainframes are still major players in cloud strategies. Why throw away something that birthed the term virtual machine.

@ChrisMatthew79 IBM has a mainframe cloud offering - DM me if you want more info - no sales pitches here...

@StevenDickens3 Thanks Steve! I figured that was in @IBM's wheelhouse.