... I think it's pretty clear they are nearly out of the running now, it's only a matter of how long they can keep taking the losses. No offense but the gameplay really isn't there.
X : What about kubernetes?
Me : Winning a battle by losing the war. The high ground you wanted to own is serverless but that fight is almost over. Amazon won etc.
X : I think you're wrong.
Me : Everyone tells me I'm wrong ...
Me : I had 100+ execs telling me I was wrong in a meeting and the future was Virtual Data Centres. Then it was VCE and vBlock. Then it was Exalogic. Then it was OpenStack. Then it was Docker. Now it's Kubernetes ... I'm very used to people telling me I'm wrong. C'est la vie.
We had our chances to create a competitive open environment ... Eucalyptus, Cloud Scaling, Cloud Foundry but they all got absorbed or entrapped by others and sent down the wrong paths. This is a tough game and to be blunt, most execs aren't upto it. Never have been.
But there's no shame in that, it's always good to understand your limits. Google has given it a try, a really good try but ... sometimes you have to lick your wounds and know when you've been beat.
X : Where would you play today?
Me : You have to move up the stack, above the run time. Conversational programming? Something in that area.
• • •
Missing some Tweet in this thread? You can try to
force a refresh
X : How much waste is in the data centre?
Me : Difficult to know because we don't often have ways of effectively measuring it. I know one particular client who got their "critical" IT estate down from 7,200+ applications to less than 600. Most the "critical" was never used.
Given duplication, bias (custom building what is a commodity), lack of monitoring, unused assets etc then as a rule of thumb, I expect 90% of all power (base resource) to be wasted but that's not just an IT thing ... in all departments I expect this - finance, marketing, HR etc.
X : Have you ordered Starlink?
Me : Yep. Hopefully here soon.
X : Why not FttH?
Me : Fibre to the home / premise. That would have been great a few years ago but it has taken so long that the future has arrived. Don't see the point now.
X : What about latency? Bandwidth?
Me : I'll comment when I get mine but given I have to channel bond to a server in London for reasons of resilience then it can only get better and as more satellites go up it will only improve. You're talking 20-40 ms and 100Mbs ...
... of course, as compute moves into space then it wins hands down.
X : Compute in space?
Me : Yep. Where do you think it's going? Ditto manufacturing. Oh, wait ... you still think land based DCs are the future? Hmmm.
X : Thoughs on industrial 4.0?
Me : In 2006-2008, I gave numerous public talks about how we were approaching a new industrial age.
X : So, you agree?
Me : Not with the term. This is about the 8th industrial age, this idea of it being the 4th is not mindful of the past.
I also don't like this pre-event classification, hence the "about", "approaching". The causes of the change seem to driven by social media, industrialisation of pre-existing activities, access to data and the changes are vast and recently accelerated by the isolation economy,
X : What changes?
Me : Long list ... 1. SWARMING (of people and machines) 2. DISTRIBUTED AND INDIRECT LEARNING 3. DISTRIBUTION OF PROVISION (not power but provision) 4. ACCEPTANCE OF STANDARDS (identification and adoption of)
...
X : Are you a socialist?
Me : My leaning is in that direction, yes. Why?
X : But you're pro markets, pro competition?
Me : Ah, the old right wing myth that socialism is anti-individualistic and pro-collectivist? It's more nuanced than that. It's more "use appropriate methods"
... i.e. we all belong to many collectives. There is a balance of Me versus We that we need to constantly review as a society. Economic systems (markets or central planning) are just context specific tools.
Many of the problems we face as a society tend to stem from one size fits all dogma of evolving methods whether it's the centrally planning of communists or the laissez faire market of neoliberals. Both extremes fail to consider context ...
X : Thoughts on Carbon Markets.
Me : The wrong way to solve the problem.
X : Eh?
Me : They are open to gross exploitation. I don't agree with them, never have. There are better ways in my view.
X : How?
Me : Every citizen should be given a non transferable permit for Carbon emission. A permit allows for an specified emission decided annually by negotiation between countries. Citizens have the right to sell or not to sell their annual output for one year on an open market ...
... so for example. citizen in the UK might get X kg and in another country they get Y Kg depending upon agreed emission divided by population. Companies must be required to calculate and buy the annual emission required from the market ...
... I do have to now ask, how long before the No.3 in the West bows out of this game?
The game of cloud was never for the faint of heart. It needs awareness, focus, intensity and the ability to play the game at the highest level.
I know Google had doubts before - cnbc.com/2019/12/17/goo… ... and that goal of being "No. 1 or No. 2 in cloud by 2023" seems far away.
There is no shame in bowing out, sometimes you just have to accept that you're not good enough. The danger is you delude yourself and stay for too long. So, I do wonder as we close in on 2023 what Google will do. 3 yrs into that journey, there is only 20 months left.