Building Resilient Web3 Infrastructure, Interview with Pauline Shangett, Chief Strategy Officer at ChangeNOW

On this interview, we sit down with Pauline Shangett. She is the Chief Technique Officer at ChangeNOW and strategic advisor at NOWNodes, to debate her current discuss at WebX titled “Head within the Clouds: Is {Hardware} the Key to Sustainable Web3 Infrastructure?”.

Pauline shares her perspective on the altering panorama of Web3 infrastructure, the myths round cloud vs. {hardware}, and why true resilience is extra about individuals, processes, and technique than know-how alone.

Within the following, the CSO at ChangeNOW shares her insights on among the most urgent issues for Web3 and past.

Pauline, in your WebX keynote you began with a placing line: “What’s the scariest factor for a CTO? The hack? No, it’s when all the pieces fails with out warning.” Are you able to clarify what you meant by that?

Completely. Once we discuss “scary eventualities” in Web3, most individuals’s minds go straight to hacks, exploits, or malicious actors. And sure, these are terrifying. However in actuality, the conditions that shake groups to their core are sometimes a lot less complicated and extra mundane: one fireplace, one missed replace, one overloaded endpoint, and out of the blue your whole product is offline.

This isn’t a theoretical concern. We’ve seen main platforms introduced down not due to subtle cyberattacks, however due to an influence outage, a defective cable, or a misconfigured failover system. And when that occurs, you’re not simply dropping uptime. You’re dropping consumer belief, transaction quantity, and in some instances, your status.

That’s why I framed my keynote as a “actuality test.” I’m not a CTO, I don’t write code each day, however I discuss to groups, founders, infrastructure leads, and chain builders continuously. And what I’ve realized is that this: in Web3, your greatest vulnerability is commonly the one you by no means anticipated, the one outdoors your assault floor however inside your operational threat mannequin. Infrastructure isn’t attention-grabbing till it fails. After which it’s the one factor that issues.

Let’s discuss cloud. Everybody is aware of the advantages like scalability, velocity, ease of use. However in your discuss, you argued that individuals overlook the safety angle. What did you imply?

The narrative across the cloud has at all times been, “It’s simple, it’s quick, and it scales.” And that’s true. However when individuals transfer away from cloud, the most typical justification is safety. They’ll say, “I don’t need my nodes or my infrastructure managed by a centralized supplier that would censor me, minimize me off, or expose me to surveillance.”

There’s fact in that concern. Centralization on the infrastructure layer introduces dangers. However sarcastically, what I see is that when individuals ditch the cloud for “security,” they underestimate one other sort of threat fully: bodily threat.

Give it some thought: the cloud’s greatest power isn’t simply elasticity, it’s redundancy. If an AWS area goes down, there are a number of layers of fallback. Whenever you self-host your {hardware} in a single facility, you don’t have that security internet. And that’s the place individuals can get blindsided.

You gave the instance of the KakaoTalk knowledge middle fireplace in South Korea, which paralyzed whole companies, together with Upbit. Why is that case so necessary for the Web3 business?

As a result of it demonstrates one thing basic: failure doesn’t need to be malicious to be catastrophic. When a hearth broke out in only one knowledge middle in 2022, companies throughout the nation froze.

It wasn’t a hack. It wasn’t ransomware. It was smoke. But the results had been huge, customers couldn’t log in, transactions had been blocked, and the federal government needed to step in. That’s a national-level disruption brought on by a single level of failure.

In crypto, we frequently discuss “crypto winters” when it comes to market downturns. However I believe the extra urgent crypto winter is operational: wars, floods, fires, minimize cables, rolling blackouts. These aren’t “edge instances.” They’re a part of the world we reside in. And if you happen to’re not planning for them, you’re primarily playing along with your infrastructure.

So how does NOWNodes method resilience in a different way? What does “planning for failure” appear like in apply?

At NOWNodes, our philosophy could be very easy: don’t ask if one thing will go improper – ask when. As a result of it is going to. That’s the one certainty in infrastructure.

Our programs are intentionally distributed throughout a number of areas: the EU, the US, and Asia, with bodily presence in international locations like Germany, Finland, the Netherlands, the US, and Singapore. That’s not only a checkbox for compliance. It’s a technique of survivability: inserting nodes the place they will stand up to political, geographical, and technical dangers.

We additionally function on a 2N+1 structure. Meaning for each crucial part like energy, compute, community, we don’t simply have one backup. We’ve got two, plus a spare. So if one system fails, site visitors shifts immediately. If the backup additionally fails, the spare takes over. It’s a layered security internet.

And we don’t simply belief the system blindly. We run common failover simulations. We deliberately shut down programs in mirrored environments to see what breaks. We do stress checks, area checks, even assault simulations. Since you don’t need the primary time you take a look at resilience to be the second an actual disaster occurs.

For years, cloud was thought-about the cheaper possibility. However you’ve recommended that equation is altering. Are you able to stroll us by that?

5 years in the past, cloud was the plain selection. You prevented huge upfront CAPEX, you solely paid for what you used, and scaling felt easy. However that narrative has shifted dramatically.

Immediately, the “Large Three” cloud suppliers, like AWS, Google Cloud, Azure, dominate the market. And as occurs in any near-monopoly, pricing developments upward. AWS compute prices, for instance, rose by greater than 20% in only a single 12 months. Virtually 40% of corporations reported cloud payments spiking by over 25% within the final 12 months.

In the meantime, {hardware} has turn out to be extra predictable. Sure, you pay extra on day one – servers, racks, energy. However if you unfold that funding over 7 – 10 years, the economics flip. One engineer famously calculated that an $1,100 server prices about $110 monthly over a decade. Examine that with $2,000 – $7,000 monthly for equal sources within the cloud, and the mathematics speaks for itself.

And past price, {hardware} offers you freedom. You’re not restricted to the options and configurations your cloud supplier provides. You may patch, tweak, and deploy precisely how it’s good to. That stage of management will be the distinction between clean scaling and bottlenecked development.

However even with nice {hardware} or a powerful cloud setup, what in case your supplier merely fails you at a crucial second?

That’s precisely the purpose. Neither cloud nor {hardware} will prevent in case your supplier ghosts you at 3AM. Infrastructure is simply as dependable because the individuals behind it.

At NOWNodes, after we ask our purchasers why they select us, their solutions hardly ever need to do with the specifics of our servers or our structure. As a substitute, they discuss in regards to the human aspect, the truth that we reply inside minutes, that we scale seamlessly with out hidden billing surprises, that we help greater than 115 blockchains together with the much less apparent ones, and that when their RPC crashed at two within the morning, our group was there fixing it in actual time.

That’s the truth: infrastructure is as a lot about belief and responsiveness as it’s about know-how.

Let’s get into the specifics: backups, multichain attain, and help. What makes your mannequin totally different?

Backups first. Nodes fail. Chains freeze. Updates break compatibility. The actual query is: when catastrophe strikes, what state do you restore? Final week’s? Final month’s? We run geo-distributed backups in order that your “worst day” turns into a small bump, not a complete blackout.

On multichain attain, most suppliers prime out at 50–70 blockchains. Solely a handful, possibly 3-5, help 100+. We help 115+ chains, and we’re continuously including extra. Importantly, we don’t simply help the fashionable chains. For instance, we’re the one supplier providing shared-node infrastructure for among the most complicated and missed blockchains: Monero, eCash, Nano, and extra. As a result of your customers gained’t wait so that you can “possibly” help their asset. They’ll depart.

Lastly, help. We don’t consider in chatbots or limitless ticket escalations. So our purchasers get actual engineers of their Telegram or Slack. Common response time: below three minutes. Decision time: hours, not days, even for deep technical bugs. And that’s not a premium upsell. That’s our baseline.

Pricing fashions in infrastructure will be notoriously opaque. How do you method transparency?

Most RPC suppliers depend on complicated tiering, hidden throttles, and shock fees. Someday all the pieces appears wonderful, the following day you’ve crossed some invisible threshold as a result of a botnet spiked your site visitors, and out of the blue your invoice triples. Or worse, they simply minimize you off mid-transaction.

We took a really totally different method. Our pricing is evident, subscription-based, and predictable. You at all times know precisely what you’re paying for. Whenever you want extra capability, scaling is quick, pretty priced, and clear. There’s no hostage negotiation.

That predictability is likely one of the essential causes our purchasers stick with us. As a result of in Web3, uncertainty is in every single place – within the markets, in regulation, in adoption. The very last thing you need is uncertainty in your infrastructure invoice.

So again to your unique query: is {hardware} the important thing to sustainable Web3 infrastructure?

No. And neither is cloud. The actual secret is resilience.

Resilience comes from good backups, distributed programs, human-centric help, clear pricing, and true multichain attain. It’s not one thing you hire. It’s one thing you construct.

Infrastructure is boring till it’s not. Till your endpoint falls, your TVL disappears, your customers rage give up, and your logs say nothing. That’s if you understand infrastructure is greater than servers. It’s belief. It’s the silent contract between your product and the individuals who hold it alive.

Remaining ideas? What ought to Web3 groups take away out of your message?

I’d say this: cease treating infrastructure as an afterthought. It’s the bedrock of your product. You may have the most effective UI, the neatest tokenomics, and essentially the most loyal neighborhood, but when your infrastructure fails, none of that issues.

Don’t ask, “How do I lower your expenses this month?” Ask, “How do I survive the disaster I haven’t seen but?” As a result of it’s coming. The groups that can nonetheless be right here in 5 years aren’t those who minimize corners. They’re those who construct resilience from day one.