A Textual Podcast

Welcome back to Roll For Enterprise, the podcast described as the squishy heart at the centre of enterprise IT. Because all four hosts were off having fun over the holidays, we couldn’t quite figure out the logistics of getting us all online at the same time to record an audio episode – so instead we put together this textual podcast, since it worked well as an asynchronous way of bouncing ideas around last time we had trouble recording together.

In the last (audio) episode we went over the major themes of 2021, so now it’s time for our 2022 predictions. Sometimes we struggled a bit to keep the two separate while we were recording, so we simply decided to double down and list the major themes of 2021 that we discussed – because we think all of these will continue to be major features of 2022:

• Semiconductor shortages and architecture turnover
• Outages and incidents
• Security in general (attacks, ransomware, etc)
• No-code/Low-Code and the shifting definition of architects
• Mental health and the change in employment landscapes
• The Great Resignation
• The year the employees took it all back

Semiconductor shortages are an easy call; all the projections forecast these disruptions to continue into mid-2022 at the very least, even if the rest of the world returns to normal. The same goes for the architecture turnover: the shift to ARM is still underway, and this year will see the software ecosystem begin to catch up with that hardware shift. More and more Mac apps now support the M1 architecture natively, and as AWS rolls out more and more Graviton-powered instance types, the same shift is happening in server software. In both cases, the performance benefits make it more than worth while to do the work of porting software to these ARM-based architectures.

As Dominic said in the last episode, outages and incidents are pretty much inevitable as long as fallible humans are in charge of systems whose complexity is at the very limit of what we can reason about. The good news is that cloud outages are short, and software can be architected to be resilient to outages of individual availability zones or even entire cloud providers. Therefore, while there may be a temptation born of frustration to blame the big cloud providers for outages that are not your fault, overall you’re still better off relying on them and the vast resources they can throw at their systems and processes. One development we do expect is a greater insistence from customers on transparency by the big providers: what went wrong, and what will be done to prevent such a failure from recurring in the future. AWS sets a good standard in public post-mortems, for instance, and others will be expected to live up to it.

The same goes for security incidents; the complexity that leads to the possibility of a fat-fingered config causing an outage also leads to the possibility of a security breach. We are not looking at any particular step-change here, just an ongoing recognition that, especially as we all continue working from home, there is no longer any validity to the idea of a network having an "inside" and an "outside". Perimeter defence is dead; defence in depth is the only way. I do expect an increase in security issues around NFTs, which will highlight the issues of decentralised architectures – and the fact that what exists today in that space is well on its way to centralising around a small number of big players.

Is this the year of low-code and no-code? Perhaps, but probably not; it’s a slow-building wave as we get more and more components in place to make these approaches fully-integrated parts of an enterprise IT stack, as opposed to weird stuff off to the side that "isn’t really IT". Partly this shift is about platform capabilities to allow for must-have functionality such as backup, versioning, or auditing. Equally it’s a cultural shift, recognising the validity and importance of these approaches as more than "toy programming". The real Year of Low-Code will come when there is an explosion of new capabilities built on these tools, built by people other than our traditional conception of developers. Right now, what we have mainly fits into existing categories. Tableau is the poster child here, but it mainly replaces Excel rather than enabling something new. That’s not nothing, but it’s not yet an industry-shifting move either.

Finally, the factors enabling the Great Resignation are still very much with us, so their consequences will continue to play out in 2022. Right now, there is a massive imbalance in large parts of IT, with new job offers coming with salaries that are several multiples of what people are coming from. This disparity is driving massive job churn, especially because companies have not changed their retention practices significantly. If your choice is between a single-digit percentage cost-of-living increase where you are, versus perhaps a triple-digit percentage increase elsewhere, the outcome is pretty obvious. If this trend continues, companies will need to get serious about retention, in part by taking factors like mental health more seriously. As we have been saying on the podcast all along, this is not a normal time. People are stressed out, tired out, and burned out by new factors and expectations, and companies need to respond to that by changing their own expectations in return. Maybe that massive raise will be much less attractive if it comes from a company with a culture of presenteeism, requiring a gruelling commute and long hours in an office with people whose health status you are not entirely sure of. That calculus becomes even easier if the company you are currently working at shows that it cares for employees by being flexible about working hours and attentive to the factors that affect peoples’ lives outside work (their own health, caring for others, home schooling, and so on).

Perhaps we close the year with a verse, with apologies to the Bard…

If we podcasters have offended
Think but this and all is mended
That you have commuted here
While our voices filled your ears
And our odd, unhinged debate
Won't predict our world’s fate
Listeners, do not unsubscribe
Do enjoy our diatribes
And, as I am a fair Lilac
If we’ve earned your candid flack
Now, to edit themes and form
Improvement shall become our norm
Else the Mike, a liar call
Or Zack incites a verbal brawl
Lend us your ears, if we be friends
And Dominic shall restore amends

🖼️ Photo by Clay Banks on Unsplash

In lieu of a normal episode of our Roll for Enterprise podcast this week, which was thwarted by myriad technical difficulties, we are trying to take our witty banter to text. Let’s see how this goes - thank you for taking the ride with us.

Mike starts us off with a bang, picking up on the release of the new Gartner Hype Cycle for Enterprise Networking (don’t worry, the link is to The Register, you don’t need a Gartner account to read it):

Mike: The reason I don’t trust most market research is that I am pretty sure vendors are writing it …

Dominic: I wish!

Lilac: Surely not. There are many capable people inside Forrester and possibly even Gartner. But, half the job is sorting through the embellishments of the vendors they do meet.

Dominic: I think this is an important point though. Many people have the impression that analysts are "pay to play", and while there are some out there that might fit that description, most of the big reputable analyst firms that you have heard of don’t work that way. There is one sense in which it is true: as a vendor, if you are a paying client of Gartner, Forrester, or whoever, you get more time with analysts, which translates to more opportunities to make your case to them. However, even in that sort of context, the good individual analysts are the ones who will pull you up if you make some sort of wild claim and demand proof, or tell you they are not hearing that type of request from individual practitioners, or whatever. These people tend to develop a personal reputation over and above that of the firm that employs them, because they provide an extremely valuable service.

To vendors, they act as a reality check, and give us an opportunity to refine our messaging and product plans in a semi-private setting rather than having to make corrections in the harsh glare of the public market.

Meanwhile to practitioners these analysts provide a validated starting point for their own investigations. For instance, if you are building a shortlist for a vendor selection, you might use the Gartner Magic Quadrant or the Forrester Wave for that market segment to double-check that you had made sensible selections. However, once again, the individual analysts leading the compilation of a particular Wave or MQ will make a big difference to the result, bringing their own experience and biases to bear, so it’s rarely as simple as just picking the vendor that’s furthest up and to the right.

Lilac: Totally agree. That has been my experience, Dominic - though we should hear from Zack here. I have been at large vendors with deep pockets that were panned by analysts - and small vendors with minuscule budgets that were lauded. Honesty, clarity, sanity … are both more valuable and harder to come by than contract dollars.

Zack: I have to be careful how I answer this question, but there aren't any surprises. Dominic’s point about vendor briefings is valid although you typically must be a paying client to schedule inquiries. Speaking from experience, a briefing is "supposed" to be one-way communication so you can update the analysts but you can’t ask questions about market landscapes or have conversations outside of the briefing. I would be more concerned about "real-world" experience as opposed to research exclusively. There are indeed some analysts that have never stepped foot in a data center although they’re basing their experience on multiple data points such as customer interactions, typically hundreds and across multiple analysts, vendor briefings & inquiries, and hands on labs in some cases, and research, but is that sufficient to form a conclusion? It might be, but as someone who spent many nights in a data center, there is nothing that takes the place of "real-world" experience. As with anything, people should use any analysts feedback as another data point in their quest to make a decision.

Dominic: Exactly – and the Hype Cycle is a perfect example, because people have a tendency to take it as predictive, assuming that every technology will eventually emerge on the Plateau of Productivity. In actual fact though there is a Pit of Oblivion somewhere at the bottom of the Trough of Disillusionment, which is where all the once-promising tech that never emerges goes to die. What is amazing about this particular Hype Cycle graphic is that IPv6 is still on it, and still in 5-to-10-years-out category!

Lilac: or VDI! It’s always the year for VDI. It’s going to be amazing.

The analysts aren’t giving you answers. They are giving you input. Things to consider. Independent checkpoints outside your organization. This isn’t a surgical specialist giving you the best answer. It’s a real estate agent, guiding you through options.

But then.. why do vendors seem to take the word of these analysts as validation and gospel? I never understood.

Dominic: At the risk of getting excessively philosophical, the answer is the same as it is for many things in grown-up life: because even with all the flaws, this is the least bad way we have found yet that is remotely practical. Right now I am sitting on both sides of different tables – wait, that metaphor sounds wrong. Swivelling my chair back and forth between two tables? Anyway. I am both running a procurement exercise which involves comparing different vendors, and participating as a vendor in a market survey.

At the customer table, I don’t have time to evaluate every vendor in even a relatively niche market, so I use analyst opinion as one of my tools to whittle down the list. Once I got to two, I took the time to talk to each one, and also talked to current customers, started figuring out pricing models, all of that – but all of that takes a lot of time.

This is kind of what I imagine readers are doing with the reports my employers participate in: not taking them as gospel, but as one input into their selection process. Getting philosophical again, it tends to be people furthest from the process who get most excited about the results – but it’s understandable: it’s one of the few results that are uncritically good. I get weirded out by vendors who trumpet their profitability, because that’s a short step to "I’m being overcharged!". Meanwhile, being certified as top of your particular field by a (supposedly) objective observer is pretty great.

But I still see no sign of IPv6 catching on anywhere. Even the hysteria about IPv4 address space running out seems to have died down.

Recommendations

Dominic

I want to recommend the latest book by Becky Chambers, A Psalm for the Wild-Built. If you’ve read her Wayfarers books (which I also recommend), you know what you are in for, even if this is a completely different setting. If you haven’t, the dedication should give you an idea: "For anybody who could use a break". It’s a delightful little SF novella that packs a lot into its short length.

Lilac

Use the 'Organizational Bullshit Perception Scale' to Decide If You Should Quit Your Job

Thanks for sticking with us! Normal service should be resumed next week. We hadn't missed an episode yet, thanks our innovative podcast architecture, which is based on a redundant array of independent co-hosts, and while it's a shame we couldn't keep the streak going, this is at least something. Apparently 26% of podcasts only ever produce a single episode, while we got to 62 before this hiccough.

Follow the show on Twitter @Roll4Enterprise or on our LinkedIn page. Please subscribe to the show, and do send us suggestions for topics and/or guests for future episodes!

I had been thinking vaguely about starting a newsletter, but never actually got around to doing anything about it — until Twitter bought Revue and integrated it right into their product.

I wrote about why exactly I felt the need to add this new channel in the first issue of the newsletter — so why don't you head over there and sign up?

And of course you should also sign up for Roll for Enterprise, the weekly podcast about enterprise IT (and more or less adjacent topics) that I record with some co-conspirators.

🖼️ Photo by Jon Tyson on Unsplash