IT trends guide in 2016–2020. Lot of marketing, mimicking and even more “unicorpses”

From author(October 2018): Original name was “IT trends guide in 2016 …” but because I’m thinking that this article is up to date I changed it to current one to emphasise that fact.

It is hard and easy at the same time to be a prophet in marketing materials when any definition can be interpreted in any preferable way and sometimes may be completely wrong. So when we are talking about “Cloud”,”DevOps”, “BigData” or other marketing hogwash we are pretty safe in our prediction. It is incredibly hard to evaluate the results in uncertain environment.

So in short year 2016 will be no different from previous years. For most companies there will be no clear vision about their strategy because of lack situation awareness and sometimes even lack of understanding of user needs. Instead, you have a chance to hear a lot of “strategy” that will be mimicking success of other industry leaders without understanding context. So, if you hear phrases like “because Gartner says so” and “how Netflix, Uber, Cisco … etc does it?” you are working with the “right” people.

“most of 2016 will consist of endless blathering on platforms & APIs by a lot of people who haven’t got a clue.” Simon Wardley

I will rephrase James Cameron “Luck is not a skill and hope is not a strategy”.

In terms of technology it becomes even more interesting because this sandbox sometimes forgets that “Engineers are hired to create business value, not to program things”. So, very often tech people faced with the situation when “BigData” gives no value to business, “clouds” are not reducing the operation costs and “DevOps engineers” are mostly overpriced system administrators with new fancy names. In other words business donsen’t care about Docker in the same way as brokers on Wall Street don’t care which shovel was used to build the trading centre.

Before I will continue, I would like to clarify who will be interested in this “guide” and what are major players who will use the buzzwords a lot.

Target audience

As you see above, the target group of this article are mostly technical people who will be developing, implementing or maintaining new systems or practices. And this “guide” tends to settle down all these hypes to practical level with all caveats.

I’m presuming that some people already have big experience in dealing with many of these concepts. And I hope they will not hesitate to extend this article with feedback and additional recommendations.

Players

I will define three categories of businesses that will be attracted to new technologies due to different reasons.

The first category are big enterprise companies who are facing to big operational costs of their data centres, doing analytics or who needs to adapt new things to compete with others. In many cases such big companies have a lot of bureaucracy, strictly defined processes and no access to real stakeholders.

Second group of customers are companies that call themselves “startups” but in reality these are new competitors on existing market with limited budget, staff and very short “time to market”. Lean approach and MVP(Minimum Viable Product) are their religion (if not, they are doomed to failure). Biggest challenge for them is bottlenecks in delivery.

The last and the most lovely group are the unicorns. “Embrace the chaos “ is probably the best description for them. The common characteristics for them are a lot of investment money, a lot of marketing and very often not a lot of common sense. Real problems of unicorns are mainly hiding not in technology but in environment and culture.

“And I blame unicorns. Not the successful companies themselves but the entire bullshit culture of swash-buckling startups who define themselves by hitting some magical $1 billion valuation number and the financiers who back them irrespective of metrics that justify it. Unicorn has become part of our lexicon in a sickening way and will no doubt become part of the history we tell about how things got so out of control again. 10 years from now people will be embarrassed to say unicorn.” Mark Suster, Upfront Ventures

Probably the best term for them is “unicorpses”. It is often difficult or even unreal to predict anything for them but exactly this group will be using all these buzzwords in their presentations and meetings on daily basis.

Back to the idea …

DevOps

Few years ago this pretty effective methodology in software development and operation cause some hype. But right now, without proper context, in most cases this is just a title of average operation engineer (administrator) who is able to write few lines of code using ruby or python instead of bash. This is perfect example of a case when something become malignant because of marketing nonsense.

As a result, “one size fits all” approach will start hitting you very hard. In case of enterprise with properly set ITIL practices, DevOps will be obliterated without any chances. In case of “unicorpeses” all depends on culture and mostly will relay at developer team rather their operation team.

Most useful implementation you will find in projects is the one that relies on SaaS solution as provider. “DevOps” will be very useful for small and distributed teams when leveraged in combination with other agile methodologies that already used by these team. And this is probably the second group of customers.

There are no losers here because most of the time business already have operation team. Renaming them and adding some toolset can improve operation cost slightly but initial capitalisation cost can spoil everything.

In short, get used to call your old admin as “DevOps engineer” and don’t bother. If “DevOps” approach works for you already then be ready to read ITIL v3 as soon as your company become bigger that “two pizza team” and you will need book opera hall for company events.

Cloud

Probably most known trend outside IT but still not defined very well. Actual “Cloud” itself is a pure marketing as other old trend “Web 2.0” and Cloud Computing are correct term to use.

In short:

1. “War” of IaaS providers ended several years ago with Amazon as the winner. Deal with it.
2. “By 2020, the game is not only all over (last chance saloon was back in 2012) but we start chalking up the casualties.”
3. “There will be some cloud players left — AWS will dominate followed by MSFT and then Google and a player like Alibaba. There’ll be some jostling for position and geographic advantages.”

Other topics to discuss is “private/hybrid clouds” and specific niche player called OpenStack.

Consider private cloud as private datacenter without a lot benefits. This market shrinks and it is very crowded. All big hardware providers are already there and there’s no space left for new players who can’t beat other by price. No wonder why IBM sold all server business to Lenovo. If you take a look at Dell, HP, EMC etc you will see what problems they have and I see no signs that their situation tend to have a better future.

You can reference to OpenStack as solution to fix that but as public cloud provider it lost its chance in 2012. Market for OpenStack implementation is very limited and will be only shrinking in the future.

In other words, if you are not in the list of top contributors to OpenStack, then don’t even bother.

The winners here are AWS, Azure, GCE(I think they will be renamed in future) and companies who are building solutions on top of them.

Losers are the major contributors to OpenStack who are still selling expensive hardware and believe that “war” is not over yet.

Microservices

Probably the biggest hypes of 2015 were Docker and “microservices”. This are really good components of the company/product ecosystem. Period.

And this is actually the problem of this movement both today and in the future. Without full ecosystem this valuable component gives no benefits for most companies. They simply don’t needed it. So many enterprise companies and startups will be using these technologies as part of release and configuration management but not as ideology. Conway’s Law is probably perfect explanation why it will be happening.

Tech companies will all be moving in the same direction:

“With all credit to Docker,” said Pivotal’s James Watters, “they are the PDF of the container world. They are the image format. And we think that we are Adobe Photoshop or Illustrator in that world — the most advanced, most sophisticated enterprise way of manipulating and managing those image formats. I think those are two distinct markets. Docker has overwhelmingly won the PDF format war, and I think there’s a real discussion over what are the more sophisticated container systems to manage it.”

The winners here definitely are CoreOS,Kubernetes and HashiCorp.

There no obvious losers here for the same reason as in case of DevOps movement. There no visible changes for final consumer. That is the question of architecture, development and operation only. The existence of REST API at Facebook make no difference for average user so Docker has even less influence.

Big Data

In case of “Big Data” product and market is better to be ready to see big battle between major providers including Cloudera, Hortonworks and MapR. During the “war”, they will dismiss utility services as “good for testing/dev but not for production”.

Reality is that most of providers receive a lot of investments money but they are not profitable yet. So best advise in this case will be to “Start preparing operation team for supporting their product” and buy a lot of popcorn.

Potentially Mesosphere could be decent competitor at this playground but their positioning as “Data Centre Operation System” and focus on “microservices” can be suicide step. As a result, they will shift into data processing domain or will be acquired (wild guess).

Finally, “Big Data” is pretty expensive “sport”, so not so many companies can afford it. In most cases it will be big enterprises that try to replace old expensive data warehouses or extend their functionality.

Others will be more pragmatic. It makes no sense to build F1 car if you just want to have a fun during weekends. Simply rent a Ferrari. And that approach is valid not only for small companies:

“We don’t have the resources internally, we don’t have the knowledge internally to build, deploy and maintain the type of tech required to leverage this info.” Coca-Cola CTO Alan Boehme

Attempt to play on log management market will fail (is failing) because of tough competition with specialised vendors as Elastic, Greylog, Fluentd and of course Splunk.

As a result, the winners in short term are the groups that will provides decent operation for existing systems. In long term the winners will be AWS, Microsoft, Google and players who will be providing “Big Data as Service”.

Internet of Things

Final buzzword on the list is interesting for me because of the type of activity it represents.

”… unfortunately IoT requires a different set of practices (from design to construction), a different set of techniques and a mix of attitude from “pioneer” to “settler”. The underlying components might be quite commodity but what is being built with these is often a process of discovery and exploration. Though there are common lessons, there’s a very different mindset and value chain relationships to IoT which is built from experience. What I’m saying is Physical + Digital is not the same as Digital.” Simon Wardley

To simplify, this domain is totally chaotic and based on a few concepts. So, if you plan to invest into education of people with expectancy that in 5–8 years they will justify your credits, then that a wise move. Otherwise it is better to be ready to burn budgets without valuable results.

I have discussion about lack of innovation and creativity in China but I’m not sure that Silicon Valley have that either. So do not hesitate to put your money into educational program in your local university instead of trying to build some IoT “product”.

Conclusions

Your experience can be different and some of things can be controversial, so don’t use this guide blindly. Common sense, situation awareness and understanding of user needs are the key components of this game. Technology need time to prove their validity and usefulness. Your profit is one of the best KPIs for your decision.

Finally, keep in mind that management in IT is decades behind even compering to automotive industry so do not hesitate to use experience from other industries.