HNHacker News
TopNewBestAskShowJobs

gz5

2,695 karma · joined January 2, 2013

submissionscomments
gz5··on Firewall rules: not as secure as you think
>fully block all internet access to endpoints you don't fully control.

Not that all risk can be eliminated, but this simplifies management while reducing the attack surface area by orders of magnitude.

The good news is companies are increasingly doing it now that technology has finally caught up - now that implementing a private* network with each vendor (or a private extranet across all vendors) is actually viable and sensible.

* Usually a software-only, zero implicit trust overlay network

gz5··on Nvidia Announces Financial Results for Second Quarter Fiscal 2025
Forget about the market reaction for a minutes, which is always difficult to interpret. More holistically, aren't we underestimating their dev ecosystems (CUDA etc.)?

Think of the dev ecosystem effects in all the various permutations, e.g. from Windows or Apple to AWS to more granular like PyTorch.

Those advantages aren't easily eliminated just by more competitors or even 'better' competitors.

gz5··on Show HN: Ruroco – like port knocking, but better
Nice. The deadline argument concept is smart and not in many other implementations.

It seems there are two sides of the spectrum for secure SSH access:

+ Relatively infrequent access by limited # of people to servers which are not top targets for attacks. Solutions like the one above are great for this.

+ More frequent, more users, more sensitive servers. Close all the inbound ports, permanently. Example: https://github.com/openziti-test-kitchen/zssh (or with an integrated OIDC like KeyCloak - https://youtu.be/NZJtzSoS_g0?si=Qg6p6Hdkaq1ahefg)

gz5··on CrowdStrike Official RCA is now out [pdf]
Note: this was distributed to their customers today
gz5··on Just disconnect the internet
we can use the Internet without being used by the Internet.

an open source example: https://blog.openziti.io/no-listening-ports

gz5··on Google Distributed Cloud air-gapped appliance
Useful for a truly never-connected 'island' (meaning it never needs to speak to the outside world).

However, even some of the use cases they cite rarely exist on a never-connected island, e.g. industrial automation and transportation.

So, to be broadly applicable, it needs to be secure by design for connected use cases as well, even if those connections are considered to be ephemeral (e.g. remote management, periodic telemetry, metadata sharing, etc.).

gz5··on CrowdStrike Update: Windows Bluescreen and Boot Loops
Seems CS themselves may have been hacked? For example, seems unlikely that both:

1. CS normally pushes global updates to entire user base simultaneously?

2. This made it through their testing. Not only 'just' QA but likely CS employees internally run a version or two ahead of their customer base?

Just speculation - folks who know either answer can validate or debunk.

gz5··on AT&T says criminals stole phone records of 'nearly all' customers in data breach
yep was trying to avoid word which carry varying connotations, e.g. vpn or zero trust.

zero implicit trust is likely the best term? you have to trust something, but enforce (and therefore trust) strong (not network based) identity, authN and authZ. this can be done anywhere via a software-only overlay.

a litmus test is server iptables (to use an example) looks like: iptables -P INPUT DROP iptables -P FORWARD DROP

and the only route outbound from the server is to the private overlay on one port, and that server still can't make those connections unless it is strongly identified and authenticated, and the overlay will not connect the client and server unless they are both authorized to communicate for that particular service(1)

(1)so for example if there is a zero day causing the 'server' to try to communicate with some_IP then the private overlay will not accept the connection, even though it is coming from the server

gz5··on AT&T says criminals stole phone records of 'nearly all' customers in data breach
Agree, good point, the overlay needs to do strong identity, authN, authZ.

The critical part the overlay adds to traditional auth is making the server unreachable from the underlay networks, reducing attack surface by billions. Meaning:

+ Let's say the server did have good auth, but there was a bug, misconfig, zero day, etc. (one of the myriads of proximate causes).

+ Since the server is available on the underlay network, that vulnerability can be exploited by anyone on the underlay (billions Internet nodes).

+ In contrast, making the server only available on the overlay, reduces the attack surface from billions of Internet nodes to the nodes which can ID, authN and authZ (for that particular server) on the overlay.

gz5··on AT&T says criminals stole phone records of 'nearly all' customers in data breach
An attacker who gets username/pw still can't get on the overlay network (the overlay requires credentials which can't easily be stolen or compromised, e.g. a private key signed X.509 certificate).

Yes, because 99% of attacks use the underlay network to access the target and exfiltrate the data. Said the other way, an attacker didn't physically walk into a Snowflake data center, console into the right server, and walk out with all the data.

gz5··on AT&T says criminals stole phone records of 'nearly all' customers in data breach
The root cause (1) is the data store should not have been available on the underlay network. Anything connected to an underlay network is a ticking time bomb.

Any servers or admins which need to talk to the data store should instead use a private overlay (2) network.

Any users (likely just remote admins) should do the same.

(1) Same root cause as 99% of breaches and yet it is too often swept under the rug while we focus on the infinite # of proximate causes

(2) Software, not private circuits.

gz5··on Twilio confirms data breach after hackers leak 33M Authy user phone numbers
consider* putting endpoints on a private overlay network in which network access is cryptography-gated (e.g. x.509 cert based).

then, a misconfigured endpoint (or a zero day etc.) can't be exploited by any_actor_on_the_internet - actors need to first complete the provisioning process you choose to enforce to be authorized to use the private overlay.

*not one size fits all, e.g. bad option if endpoints need to accept requests from unknowns.

however, many endpoints only need to accept requests from known (identified, authenticated, authorized) endpoints, and the added friction to id/authN/authZ get use the private overlay is not a business impediment.

there is a stigma here due to the horrors of NAC on private enterprise WANs. but NAC goals can be accomplished without that baggage via internet overlays and modern cryptography.

to be clear, i am by no means advocating to abandon traditional methods of endpoint auth - this it is just another layer which recognizes that single layers are rarely airtight (e.g. what just happened to Authy and Twilio).

gz5··on Cyber Security: A pre-war reality check
>Why did it happen? Non-technical people have made choices and have optimized for stuff being cheap.

Yes and amplified by:

+ Cybersecurity 'bad actors' are decentralized and distributed. They innovate at speed, with no barriers, and share their innovation. Cybersecurity 'good actors' are centralized, proprietary and bounded.

+ Software and service providers traditionally couldn't build secure networking into their products - they had to delegate it to the consumer of the software or service for the consumer to implement as a day two bolt on. Dangerous when networking is often the largest and most vulnerable surface area.

gz5··on Novel attack against virtually all VPN apps neuters their entire purpose
Encapsulate and encrypt in the app itself, or in the browser.

App (via the openziti sdk): https://blog.openziti.io/no-listening-ports

Browser (the openziti js sdk loads on the fly): https://blog.openziti.io/introducing-openziti-browzer

Disclosure: openziti (apache v2) maintainer

gz5··on JIT WireGuard
For Chromium-based browsers, an option is to use BrowZer (built on OpenZiti, Apache 2.0). Enables you to connect into a full mesh private network (mTLS, e2e encrypted, no TLS man in middle inspection). 3 examples below with well known apps. Disclosure, I work on the project.

MSFT RDP (video):https://youtu.be/1NMrxRIowog

Private network for Grafana (video):https://youtu.be/l5ktiI-j3eg

Private network for Plex (blog post)https://blog.openziti.io/its-a-zitiful-life

Basically you decide what 'app' you want to deliver via the overlay, e.g. Grafana, Plex, RDP. For those destinations, a (one time) bootstrapping process (invisible to end user) results in your browser receiving a <script> tag which includes some configuration when the browser attempts to connect to the destination (Grafana etc). This ultimately results in the browser downloading some JavaScript and WA, and registering a service worker (the wasm contains the PKI bits).

After successful auth, your browser can then open a websocket to your private OpenZiti overlay network (distributed, OpenZiti overlay network software routers, deployed where you want them), and ultimately hit the app (which no longer needs to listen to anything other than the overlay network; becomes private).

Desktop Chrome is the most tested, followed by Android Chrome.

gz5··on A supply chain attack on PyTorch
there are 3rd party foss options (1):

1. ephemeral + zero implicit trust (2) https://blog.openziti.io/my-intern-assignment-call-a-dark-we...

2. zero implicit trust: https://github.com/openziti/ziti-webhook-action

(1) disclosure, maintainer (2) zero implicit trust in this case = no open inbound ports on underlay; need to access via app-specific overlay which requires strong identity, authN, authZ

gz5··on Ask HN: Who is hiring? (December 2023)
NetFoundry | $115k to 180k base comp | Full-time | REMOTE | USA | DevRel Leader

You are responsible for building the OpenZiti developer community. This is initially an IC role; you will then build out a DevRel team.

The open source OpenZiti platform enables developers to embed zero trust, full mesh networking into their solutions, as software. Example:

https://github.com/openziti-test-kitchen/go-http

OpenZiti delivers billions of zero trust sessions per year for leaders around the world. NetFoundry originated OpenZiti, is the maintainer and sells CloudZiti SaaS (hosted OpenZiti). NetFoundry is a remote-first company.

You are a good candidate if you love to learn about the problems, opportunities, constraints, hopes and fears of the developer, operations and security communities which you will serve.

More info:

https://netfoundry.io/careers/devrel-leader/

gz5··on OpenAI's chaos does not add up
Ironically (in an AI context), actions driven by human sentience has to be the #1 factor enabling this.

I do think there are sub-factors, e.g. California legislation against non-compete and non-solicitation enabling Microsoft to (apparently) offer to hire dozens of OpenAI employees.

gz5··on New Microsoft Exchange zero-days allow RCE, data theft attacks
well said and i agree. i believe we are talking about different layers. your point - essentially assume your network is always breached - absolutely.

and don't you make that 'battle hardening' simpler and more effective by reducing the attack surface? e.g. by taking your servers off the internet (meaning your inbound firewall rules become deny all inbound (even 443)). so enforce (strong) auth outside your dmz, before allowing sessions on your network (or overlay network), even for APIs, B2B etc (otherwise your fw has exceptions).

and, yes, when that gets compromised, the attacker now needs to deal with the next set of 'battle hardened' layers.

meaning, shouldn't reducing attack surface and battle hardened services be an 'and', not an 'or'?

gz5··on New Microsoft Exchange zero-days allow RCE, data theft attacks
this is where zero trust has steered us wrong, inadvertently.

absolutely perimeter security based on weak auth (being 'on' the WAN) is insufficient.

but improving internal security doesn't address all internet-based attacks - which are the vast majority. in those cases, auth before connect is a good practice - but the auth (authentication and authorization) itself needs to be strong, and part of a multi-layered approach (the WAF etc are actually able to be simplified and do their job better when they don't need to filter the entire internet).

gz5··on New Microsoft Exchange zero-days allow RCE, data theft attacks
good point. i simply meant that the vulnerability can be exploited from the network (with no (initial) root access to the machine) and so almost all of them are.
gz5··on New Microsoft Exchange zero-days allow RCE, data theft attacks
agree, but shouldn't we try to further reduce the attack surface? e.g. the 'server' only listens on networks which force 'clients' to authorize before those clients are given access to that network (not always possible, but often is). sometimes for example this can be an auth before connect overlay network extended to the host such that it is only listening on localhost.

and, agree, the clients can be compromised - loved ones can be scammed as you said - but your loved ones are a far smaller attack surface than being exposed to any attack on the internet (which too often can find its way into the dmz until day two security and L7 authorization tries to identify and terminate the rogue L3 connections).

gz5··on New Microsoft Exchange zero-days allow RCE, data theft attacks
Look at any major CVE and you will almost always see "...that attackers can exploit remotely".

It is logical - 9x% of large cyber-attacks are done digitally, not with physical proximity to the target.

Yet, we often focus on the vulnerability (zero-day, misconfiguration, business logic gap, etc.), rather than the exploit method (the network). Almost implicitly taking it for granted that the server (e.g. Exchange) needs to be exposed to the network in order to do its job.

Given the impact, shouldn't we double down on methods which enable servers to do their job without 'listening' to the network?

gz5··on Ask HN: Who is hiring? (October 2023)
NetFoundry | $115k to 180k base comp | Full-time | REMOTE | USA | DevRel Leader

You are responsible for supporting the OpenZiti developer community. This is initially an IC role; you will then build out a DevRel team (our engineers do our DevRel today).

You enjoy the opportunities and pains of working at a startup, and you love to learn about the problems, opportunities, constraints, hopes and fears of the developer, operations and security communities which you will serve.

The open source OpenZiti platform enables developers to embed zero trust networking into their solutions, as software. OpenZiti delivers billions of zero trust sessions per year for leaders around the world.

NetFoundry originated OpenZiti, is the maintainer and sells CloudZiti SaaS (hosted OpenZiti). NetFoundry is a remote-first company.

More info:

https://netfoundry.io/careers/devrel-leader/

gz5··on Update on KDP Title Creation Limits
I think we will see tidal waves of 'not-so-good' AI-generated content. Not that AI can't generate or help generate 'good' content, but it will be faster and cheaper to generate 'not-so-good'.

These waves will mainly be in places in which we are the product. And those waves could make those places close to uninhabitable for folks who don't want to slosh through the waves of noise to find the signal.

And in turn that perhaps enables a stronger business model for high quality content islands (regardless of how the content is generated) - e.g. we will be more willing to pay directly for high quality content with dollars instead of time.

In that scenario, AI could be a_good_thing in helping to spin a flywheel for high quality content.

gz5··on Show HN: Every Breath You Take – Heart Rate Variability Training
I have found HRV to be remarkably accurate at 'predicting' illness (seeming to indicate that HRV is impacted before I notice symptoms). Do others see the same?

I do need to account for 3 other variables:

+ did i drink alcohol that night

+ did i eat a large meal right before bed

+ did i go on a long run (20+ minutes) that day

Those 3 significantly decrease my HRV (alcohol the most). On the other side, not eating within ~6 hours of bedtime significantly increases my HRV, and moderate exercise increases my HRV.

Curious what others are seeing? I have measured with watches and rings - both which only measure at night.

gz5··on The Tyranny of the Marginal User
And the tyranny extends because most of the money goes to those who can (theoretically) get geometric growth - get lots of us during our Marl moments.

If the pendulum swings the other way - meaning we each have less Marl moments - more of us willing to pay for an experience with dollars than time, then expect more OKCupids to start and to persist. Theoretically without venture money.

I think LLMs may help swing the pendulum in the OkCupid direction but in two very different ways - curious how others see it.

1. Lots more Marl courting apps developed, faster. Causing more fatigue. Less time to go to any single Marl app; reluctance to try new ones. Push pendulum away from Marl apps - each of us cuts our Marl moments.

2. Ability to develop OkCupid apps with less investment and target smaller circles (which often means i can deliver a better experience to that small circle, and not need VC funding/geometric growth to keep the lights on). Pull pendulum towards OKCupid apps.

gz5··on The Silicon Valley elite who want to build a city from scratch
Telosa is a similar idea on a larger scale. They are targeting 5M people population by 2050. Original idea was the citizens all have equity in the value of the combined land.

https://en.wikipedia.org/wiki/Telosa

https://cityoftelosa.com/

gz5··on Verifying your identity on Twitter will require taking a selfie
linkedin also just made this an option (and are soft pushing it). payment apps likely started the trend - stripe, venmo - but that was driven (afaik) by kyc.

maybe x/twitter, linkedin etc. plan to add financial transactions and will need to be kyc compliant - so this is just a first step?

or is there a new 'driver' for this model to cause companies like these to prioritize this 'feature'? is there really that much destructive fraud or impersonation today?

gz5··on Kubernetes Exposed: One YAML Away from Disaster
and initiatives to make devs/ops/systems use vpn, bastion, static permitted IPs, etc. are often DOA due to their impacts on automation and speed.

a foss alternative is kubeztl - it takes kubectl off the networks - no public IP at all - but w/o users needing agents: https://github.com/openziti-test-kitchen/kubeztl/tree/main

disclosure: i am a maintainer and the software overlay in the middle (helps enforce outbound-only, pre-authorized connects only) needs to be managed (self-hosted foss or hosted saas), so there are still trade-offs.

← PreviousPage 3 of 14Next →