This video is fantastic. Funny, smart satire with real commentary. Worth watching.
This video is fantastic. Funny, smart satire with real commentary. Worth watching.
Enterprise IT is willing to pay high prices and deliver large profit margins to suppliers. Stop whining about the price.
Lets consider the evidence:
In technology, there are no penalties for spending up big in the name of downtime. You sit on your cheap chair, look at your cheap desk in a rented building, check your payslip (you get paid the least amount possible) and get ready to buy the most expensive technology because its the “right thing to do”.
Tonight, as you drive home in your well-priced car, think to yourself “Do I really need buy the most expensive IT products on the market?”
At least, stop pretending that you are getting a “good deal” or the “best price” and then complaining that technology is expensive. Thats all on you for not buying the cheaper option every now and then.
Takeaway: DevOps methods are first rate staff, second class assets. Helluva difference from Enterprise IT.
When I’m working with people on IT Infrastructure, the existing habits (learned responses?) of IT managers is to invest heavily in high cost, high quality systems from known vendors that have high maintenance costs. These first class assets are expected to have low cost of operations & high reliability.
Cost can be saved by hiring second rate staff who don’t understand technology but know enough to operate the systems to a lesser or greater degree.
Proof point: compare the demonstrated skills of big cloud providers using low cost, no support, no maintenance assets.
Working with DevOps, its quickly becomes clear that formula is reversed but comes out with the same result. In DevOps, hiring first class staff is critical. Staff who are high cost, high quality who are able to self-motivate, self-organise and adapt to multiple requirements, languages and evolving changes in their platforms. Those staff are also ‘high maintenance’ needed time to retrain, retool and replace their skills constantly.
In a DevOps environment, the use of high quality, high cost systems is not needed. Customers can choose to run on low-cost, second-class assets.
Sure you can. The question is whether this level of spend/investment will produce a return. Having high staff costs AND high infrastructure costs (don’t forget the maintenance/service contracts here) doesn’t seem to make a lot of sense.
Thats why public clouds use low cost infrastructure.
The future of networking, and IT Infrastructure more generally, will be driven by machine to machine interactions that are the heart of Internet of Things.
The current generation of networking products, protocols and architecure are driven by the fundamental assumption that a
human meat puppet will initiate connection from a client to a server. Consider:
Traffic is generated when the meat puppet engages with content. At this point, network traffic is generated, server resources are consumed, logs & analytics are monitoring etc etc along the food chain.
Today, the IT market is driven by the demand of consumers because this is the largest market using IT products. Enterprise IT is much smaller market.
The term Internet of Things is a catchall term for the next transition in technology. I consider that “Things” replace “meat puppet” as the source of the traffic. These “Things” will use the “Internet” as the network at first. I can go with “Internet of Things” (IOT) as a workable definition of a movement or a comprehensive market trend while being nuanced enough to recognise the lack of a concrete definition.
Think carefully about the interaction model of IOT. Your personal health device works without your attention. It collects data as programmed and ships it to a remote application. No human required.
Things vary according to the purpose. There are ‘personal things’ like Fitbits and Apple Watches. There are ‘building things’ like locks, fire systems, temperature and airflow sensors, elecrtrical current monitors. There are ‘factory things’ where installed production plants will retrofit a new generation of sensors for visibility and analytics.
‘Office Things’ will monitor meeting room and desk utilisation, daily attendance, common pathways so that employers can track office efficiency and utilisation. Throw in some security things as well.
‘Infrastructure Things’ are upgrades to power lines, water reservoirs, natural gas flows, sewage status and more.
You get the idea.
All of these ‘Things’ will generate network traffic and use servers for storage/compute. No humans will trigger activities, it will all be controlled by algorithms in the devices or cloud applications that configure them.
This creates new patterns of resource consumption across our infrastructures. There is no “peak time” for a device. There might be “critical service” on emergency things like fire sensors. There could be zero service levels on “meeting room vacancy” thing (who cares if miss a few minutes of data).
There is a key enabler for IOT and its the “software defined thing”. Everything that has been achieved in the last five years for software defined infrastructure has paved the way for the automated/orchestration of the ‘things’.
Consider how many ‘things’ are going to be deployed in the next two decades. Today, people think that three devices per person is a lot. Think about a building
The machine revolution is well on its way.
Big companies are putting billions into thing-filled jelly wrestling pits so that they can attract partnershsips with companies who have “customer relationships”:
Infrastructure technology that can support the IOT products of these companies is a whole new market. Its a market that is vastly bigger than Enterprise IT with fewer customers and lower overheads because its a platform.
This week Cisco announce the USD$1.4B acquisition of Jasper to gain an established IOT platform that has established customers and relationships. Thats one of the biggest acquistions for Cisco since SourceFire four or five years ago and spending US dollars doesn’t happen often. Evidence that growth is anticipated in this market.
Just because I share my opinion publicly doesn’t make me an analyst. I’d rather be known as an Activist Customer than a Citizen Analyst since there is widespread perception that industry analysts are biased or easily influenced.
The Geek Whisperers podcast discusses how social media is changing the public face of vendor / customer relations and has been creating names for all the things in social media. I enjoy listening to the show to gain insight into the machinations and manipulations that PR/Marketing people go through as they find the best way to communicate with customers of their employers.
A recent discussion came up with a name for bloggers as “Citizen Analysts”. I find this term uncomfortable because of associations with analyst firms and which I continue to have conflicting opinions about. In short, while analysts firms are taking money from the vendors they are open to influence by vendors. Analyst firms can’t survive without revenue from vendors so this fact isn’t going to change anytime soon.
But you shouldn’t criticise unless you have a better suggestion which I’m putting forward here.
I would point out that I have never worked on the ‘sell side’ as a vendor or distributor. I’ve never worked at an analyst firm. I have only worked for resellers and customers. I may yet work for a vendor or an analyst firm in the future and have my opinion changed.
I have limited knowledge about the “industry analyst” business but I certainly know how to perform industry analysis from a customer perspective. When I’m working as part of the architecture and strategy team on a customer site I will be tasked with analysing vendor products — their strategy, their financial position, the competitive pressures in the markets and so on. I will write reports and reviews to be submitted to committees and team reviews. I’ve seen analyst reports from big firms and the results are similar.
I take a keen interest in the movements in the networking industry. I invest time in observing what each vendor is doing, how the products evolve and look for patterns that are distinctive to each company. After a while of doing it’s reasonably simple to base your customer strategy and architecture on those patterns.
Some people call this “predictions” or “industry analysis” but I call it “work”.
Over time, it’s become clear that anyone who blogs will be have their opinions viewed and “heard” by vendor employees. I’ve had contact from vendor employees at all levels of their businesses – from engineers to the senior executive team as a result of articles I’ve written. It has been a mostly positive experience as well as periods negative engagement in more recent times. My impact (or influence) seems to be similar to large customers whose purchasing power gives their requests a certain gravity that can influence new features or products. This is a form of customer activism using purchasing power to create change in the vendor roadmaps or request a vendor to look at some problem or another.
My form of customer activism is to share openly and independently.
Over time, I’ve met many people who have taken the time to write their experiences on the web. The common themes are about sharing, to improve, get feedback and to create change. Those terms are also related to civil activism related to politics. I’m not suggesting that everyone rush out and starting agitating like radical protestors, that would be pointless. But I can bring positive and negative aspects of being a customer to the attention of vendors who might otherwise not hear about these issues. Sometimes they can be addressed, sometimes not.
But I prefer to the term “Activist Customer” to “Citizen Analyst” because analysts are commonly regarded as biased or influenced. What are your thoughts ?