A cloud computing eddy current in the Evolution axis?

Apologies for the random observation here. Don’t know if anyone caught the recent a16z post on repatriation of cloud services as a cost containment measure:

Coming at this phenomenon with a Wardley Map lens, you might even see this as an example of “evolution in reverse”. So it begs a few questions, like if the de-commoditization of cloud computing a measure of altering the value chain of services delivered?

Not so from the article, however. This is framed as a pure cost management play and not one of bringing value closer to the end customer or creating space for creative innovation and opportunity by unbundling, for example. Or maybe repatriation is a form of unbundling to unlock value that can be better cost-optimized for a single large customer? Making it worth moving from utility back to infrastructure.

I’m sure other patterns like this exist “in the wild”. Just curious if other experienced mappers had different interpretations of what would be at play in this scenario… or if there are good, similar examples to cite.

@greg what’s an “eddy”? I suspect you mean this, but figured it would be worth asking :sweat_smile:

Now, on the topic, I can offer two thoughts:

  1. were they trying to exploit the internet adage that to get an answer to a question you need to post the wrong answer? In other words, were they trying to crowdsource the solution to (possibly, theirs only) problem. This would ask of us to take a different, more critical, perspective on its premises and conclusions. This paragraph seems to confirm this (humorous attempt at) interpretation

The point of this post isn’t to argue for repatriation, though; that’s an incredibly complex decision with broad implications that vary company by company. Rather, we take an initial step in understanding just how much market cap is being suppressed by the cloud, so we can help inform the decision-making framework on managing infrastructure as companies scale.

  1. I like the points raised by Corey Quinn on this thread shared on twitter a few days back: https://twitter.com/QuinnyPig/status/1398006284209688578 which, after a fourth read, reinforce even more my belief about item 1 that I just wrote above :laughing:

Tl;DR: there’s no eddy (as in a circling back to on-prem) at all

Thanks, @umberto.nicoletti . :smiley: You got it… by eddy I mean an eddy current. As in something that may spuriously seem to reverse its normal direction of flow, but not so much as an aggregate momentum forward. Apologies for being a bit obscure there. :slight_smile:

Thanks for the interesting thoughts…

  1. I don’t think so. Venture Capital firms hate being wrong. Or even worse, they hate being perceived as being wrong. So it’s a bit of a stretch to imagine they would put themselves out there with the deliberate attention to potentially tarnish their thought leadership façade in order to crowdsource answers they were unclear about.

  2. Cloud governance seems to be at the root of the inefficiencies. A lot of it seems rooted in a lack of systems of awareness of services used, services UNUSED, and collective cruft that accumulates in a cloud datacenter footprint with few able or aware of how to sort out the tragedy of the commons within.

Saying there is no eddy here would suggest that this is two steps forwards but one step back for those who might better take ownership of a utility in-house. But that would only make sense at scale.

It’s rare to find situations where someone takes a utility and in-sources it with the assumptions that they could operate it better internally with a customer base of one … unless they had some strategic advantages they could enable by doing so. Cost management alone doesn’t seem sufficient to me.

Sure, I could save money by cutting my own hair and changing my own motor oil on my car instead of taking it into the shop. But WHY?

If I look at that post from a (naive, maybe) VC point of view here’s what I would think:

“we gave X bags of money to that startup and they’re burning through it by spending it all on cloud. They could run out of money less quickly if they used it more judiciously instead, i.e. by running some workloads on-prem”

Maybe they see this problem that we don’t get to see?

to potentially tarnish their thought leadership façade

(warning: sarcasm ahead) I thought it’s a gambling game, TBH I’d never thought of thinking of them as “thought leaders” :sweat_smile:

I do like the merits of your hypothesis here … that it could be a way for a VC to say, “We told them they could be more profitable if they followed our advice, and here’s the evidence of ours that they’re ignoring.” Pretty passive-aggressive, but possible. I’m not convinced it’s likely though.

Oh, but VCs are trying to compete these days as content marketers. It’s strange, but true in a lot of circles:

and here’s even a citation about a16z’s content marketing strategy in Forbes from 2013:

Reads like using dialogue mapping. Solving wicked problems by proposing solutions and evaluating the pros and cons.

My own view (and I don’t want to read the article before I comment) is that repatriation is a good strategy in times of uncertainty. Take over as much of the “supply chain” as possible.

A lot of companies have found their cloud costs ballooning (that’s a metaphor, just like “eddy”) and there is a distinct move to edge computing.

Now the current edge computing is offered by cloud suppliers who have an extensive (and expensive) content distribution network (CDN). The customer pays again to get better response.

Yet a multinational company would benefit by hosting in the countries or locations where they have a presence and demand.

I see this repatriation as part of a transformation from old cloud (centralised services being offered from surplus computing capacity) to new cloud which is distributed and eventually peer to peer.

Nothing stays the same and I may have the wrong end of the stick (another metaphor).

I agree with your opinion. VC’s are herd animals.

As someone who’s (still) running most of their services on-prem and evaluating/planning a move to the cloud I have to admit (and agree with you AFAIU) the cost-saving reasons look very weak.

Perhaps, as Corey put it, it made sense for Dropbox as they had a single well understood workload (storing files). But for “general” computing cases, I really think it’s hard to justify on-prem on cost-reasons only.
And even then, there’s a good chance that costs are just being inaccurate (to put it mildy) for on-prem.

Using a Wardley map approach to analyze it: cost-optimization or cloud governance (whatever you call it) is a novelty practice that’s emergency on top of commoditized compute services. This, I think, makes very much sense.
However, if that’s the idea of the article, I’d venture to say it was not expressed very clearly. Maybe it needs a second read, but I don’t have time for it :sweat_smile:

Read the article, typical VC view; all numbers and metrics. Not a mention of change of cloud infrastructure (edge). I think this article is for investors.

Here’s a quote from the article:

For those who have not planned in advance, the necessary rewriting seems SO impractical as to be impossible; any such undertaking requires a strong infrastructure team that may not be in place.

A blind spot, most large multinational organisations will outsource complex developments.

Only companies with storage as core (LinkedIn) will have internal development teams.

Repatriation is a competitive advantage when resource flexibility is no longer important.

I would recommend all new development should be on cloud and a decision made if the infrastructure comes in-house.


I second this. Corey also mentions that cost-savings go after time-to-market. I imagine that you can start thinking about taking things off the cloud when your offering stopped changing, and few % of efficiency means a lot in absolute numbers.