Thursday, September 5, 2019

The IT Toolbox #007 - Definitions #3 - drink the SHIFDX

Drinking the sugar high inducing flavored drink mix.

Because of and © we should probably call it SHIFDX.



It is the tendency by people to indulge in exuberance, especially when it comes to marketing information.

In general, it's an important part of the technology industry, because it gets the word out about new technologies.

In specific, it leads to very poor business decision making.

Consider every decision made in the early hype of a technology and where that technology fits in the solution stack today.

Odds are:

     A) it made its way along an evolutionary path and the hype died off - we actually figured out what it was really good for

     B) it isn't what the marketing perception projected - it does something in the realm of what marketing said

     C) it is being impacted by a new technology/integration/abstraction - someone figured out an enhancement or better way of making/doing

     D) it is either WAY cheaper or WAY more expensive - yeah, go figure you have to experience it to actually understand what it cost

     E) except for the companies that created the product, the market changing benefit has ... changed

My suggestion, if you're going to "drink the SHIFDX", do it in moderation and make sure you hand your keys to someone until you're sure you can drive again.

Thursday, August 29, 2019

The IT Toolbox #006 - Multi-Cloud Strategy

A multi-cloud strategy is half-way between two types of thinking.

It's also not thought of as an optimal technical solution for a business problem.

But, it's the only way for late majority users to adopt the new technology lifecycle.  It's also the elephant in the room.

Let's start with Roger' bell curve:

Figure 1.  Roger' bell curve.  credit:Wikipedia.org

Conceptually, the businesses that are going to use a public cloud strategy effectively are the ones that are arguably already using it.  They were the innovators, AWS, Google, Azure, etc.  What's interesting, in the case of the first two, the lifecycle that lead to public cloud was an arbitrage of their excess capacity.  A way to wring out more value from already in place infrastructure.

The "earlies" saw this as a starting point for their needs.  Through consumption of a defined model of delivery, their use cases fit the new lifecycle model.  At any relative scale, public cloud was a way to reduce the business decision of build vs consume.  Netflix is an incredibly good example of an "early", using the new lifecycle model literally made it possible for them to concentrate their effort more specifically on developing their product.

The late majority has a different business problem than the innovators and the "earlies."

They desperately want to be able to take advantage of the advances they perceive in the technology change:

Figure 2.  Enterprise Virtualization vs Public Cloud - Link

I've previously described the state of advancement of enterprise virtualization vs @swardley 's Public Cloud map using his technique (above).  

The difference under the Wardley mapped model though, is that the Use Case doesn't align cleanly.

It is also why there's a contentious argument in the minds and words of business owners.

During the peak of expectations, we hear statements like this:

     "We're moving our workloads to public cloud."

     "We'll be 100% public cloud in 2 years."

In despair, the story changes.

      "Public cloud is too expensive."

      "Public cloud is not secure."

      This is also why Workload (or Cloud) Repatriation occurs.

With enlightenment, the story changes, yet again.

     "We're implementing with hybrid cloud, so we can take advantage of new technologies and techniques."

Today, anyone that says they are moving from legacy to hybrid cloud is perceived as a laggard.

Consider this:

When a business from the "earlies" in public cloud moves to hybrid, it is a conscious business decision. It's thought of as forward thinking. 

     Consider the infrastructure necessary to deliver Netflix today.  It's not pure public cloud.  In order to meet the content delivery goals with their customers, they built a Content Delivery Network (CDN), integrated with their platform.  It is hybrid

     When the public cloud innovators and the "earlies" start to invest in on-prem workload execution, it is hybrid.

Multi-Cloud Strategy is meeting in the middle ground.

     It is a legitimate tool in the #ITToolbox.

     The immediate future is hybrid.

Monday, August 12, 2019

IoT - Wardley Maps WAR strategy


This conversation starts from Chapter 9, of WardleyMaps where it might be useful to use a map to look at the influences of predictability on a specific topic.

The next weak signal leading to predictability from Figure 106, future points of war, is what I'm looking at for Internet of Things (IOT).

I'm using a map that I've been toying with for a while.  The concept is relatively straight forward, what actions will lead to the industrialization of IoT?

We'll first start with what the map looks like for one aspect of IoT, I'm choosing data presented of some value to the user.

IOT, Wardley, Map, WardleyMap, Value, Chain, WAR
Figure 1. IoT User Visualization - Wardley Map
The second thing we should discuss is where value is derived.  Arguably, value is different to different people.  From an IoT customer's perspective, the value of the Visualization is far more visible than that of Power, Data Center or the computers it houses.

Therefore, if you're building IoT capability, the areas in the upper left are the areas you're going to want to invest in.  A simple representation of this idea on a Wardley Map looks like Figure 2.

IOT, Wardley, Map, WardleyMap, Value, Chain, WAR
Figure 2.  High, Moderate and Low Value on a Wardley Map
  The third area leads directly to pace, derived somewhat from Figure 2 and relevant because areas of attack should be at the pace required by the results.  Basically, don't lay siege to an entrenched industry, find the weak points for quick wins.

IOT, Wardley, Map, WardleyMap, Value, Chain, WAR
Figure 3.  Methods overlay on a Wardley Map
The concept is relatively simple.  Do you attack an IoT investment area in a Lean/6-Sigma area, that is already well entrenched and likely not to change rapidly, Figure 4
IOT, Wardley, Map, WardleyMap, Value, Chain, WAR
Figure 4.  Slow attack areas on a Wardley Map
 Or, do you attack an IoT investment that is in a rapidly changing area where it's using something like Agile Methods.  An area where the feedback to new method/enhancement/capability can be established in a very rapid method and is likely familiar to newer companies/organizations.  Figure 5.

IOT, Wardley, Map, WardleyMap, Value, Chain, WAR
Figure 5.  Fast attack areas on a Wardley Map
With that perspective, the WAR over IoT will be conducted, in general, in this area.  Figure 6.

IOT, Wardley, Map, WardleyMap, Value, Chain, WAR
Figure 6.  Iot WAR on Wardley Map



Tuesday, July 16, 2019

The IT Toolbox #005 - Thoughts on Cybersecurity



Define a set of cybersecurity rules.

Define an architecture (be it physical, platform and/or application).

Make sure the aforementioned rules can be applied.  (It doesn't matter if they are perfect, NONE are.)

Fix the rules or what the rules break.

For the love of all that is holy, PATCH in a reasonable amount of time.  (If you use a service provider, make it a contractual obligation and/or a Key Performance Indicator (KPI).)

Make sure there is a mechanism to verify the patches are in place.

Make sure there is a mechanism to verify FW rules are CORRECT.

Segment ALL applications.  Microsegment all unique elements of all applications.  Use SSL.

PATCH everything in a reasonable amount of time (yes, it's a repeat, but many don't hear it the first time).

Be prepared to burn down ANY exposure.  Have a plan in place in the event this must happen.

Have a reporting and notification plan in place.

When an exposure is identified (and it will be) make sure you use the reporting and notification plan.

If you EVER have to break ANY of the self imposed Cyberscurity rules, segregate and enclave to limit exposure.



Tuesday, July 9, 2019

The IT Toolbox #004 - Definitions #2



AI is not AI

AI is a marketing buzz word (ok it's an acronym (or even an Initialism), but bear with me)

Definitions of AI I've personally witnessed

     AI is stealing jobs (well, boring, repetitive, mind-numbing jobs)
     AI is Machine Learning (it can/could be, but mostly isn't...yet)
     AI is data processing (so is paper shuffling)
     AI is a Virtual Agent (actually partially true, at least the native language parts)
     AI is The Terminator (nope, that's a Movie, starring Arnold Schwarzenegger and on my must                                                  watch list)
     AI is Analytics (sorry, Analytics is Analytics, representations and uses of data)
     AI is a self driving car (hopeful a self driving car is more than AI, wishing for seats, engine, etc)
     AI is Deep Learning (might be true, how supervised is the learning?)
     AI is self service (well, ok.  Gives supermarket self checkout a new meaning.)
     AI is Robotic Process Automation (not so much as a set of response triggers, but bits could be)

If you’d like to read a thoughtful description of Artificial Intelligence, have a look at this Wikipedia article: https://en.wikipedia.org/wiki/Artificial_intelligence  (15 pages and 375 references, AND 18 disambiguation references)

So, if you're talking about Artificial Intelligence (AI), consider refining your definition and talking points to if there is (or is not) a neural network being trained in support of the mymicing of "cognitive" functions.

If not, call it what it really is rather what someone is marketing.

Wednesday, July 3, 2019

The IT Toolbox #003 - It is only a little off

IT is complicated ...

Here are some sensors:

Primary Adoption Strategy of Digital Transformation 1 
    (surveyed IT departments)

     34% - heterogeneous IT integration (basically picking parts that work for a particular purpose)
     27% - entirely public cloud
     24% - entirely private cloud
     16% - hybrid cloud

     Less than half of enterprises (surveyed) have a mature cloud adoption strategy

     12% - self report as mature
     37% - self report as somewhat mature

84% 2  of public cloud customers will repatriate some workloads to private infrastructure in 2019

Between 40% and 80% of enterprises will fail to deliver traditional workload on public cloud 3,1 

Enterprise Data Centers are closing - incorrect 4 

     Enterprise Data Center spending continues 5,6

x.86 Market growing at 19% 7 

Major x.86 vendors are growing market share and revenue 8 

There's a massive misunderstanding about the definition of the Digital Transformation end state. 9 

--  Publically available references --