An accurate title would be "IBM archives the IBM-HashiCorp Boundary Community Edition repo".
That "HashiCorp" in the title is a farce. The "Hashi" of the HashiCorp is far gone. IBM and "Corp" are redundant.
Community Editions don't make sense to the MBAs at IBM unless killing those off would create a global backlash (loss of goodwill for no positive revenue outcome is not good to an MBA. If loss of goodwill however, results in positive revenue, that's a good thing to an MBA. This is a well regarded, well studied and highly respected, revenue boosting MBA concept called "enshittification"). MBAs know that anything that doesn't directly translate into either viral growth or revenue next quarter is a waste of investor dollars. That makes anything else a risk to their own careers.
Boundary is past its viral growth stage, so the MBA needs to see revenue next quarter. Boundary fails the test there, so off with it!
HashiCorp was an innovative company founded by two engineers that built a lot of foundational, transformational infra products. Then, in a growth frenzy based off the traction of these tools they raised tremendous, vast sums of money and subsequently ~lost control of their vision~ hired MBAs, and the company went down the drain as well, or as MBAs call it - got acquired.
What remains of those tools are a BSL that focuses on maintenance operations and feature requests that strictly ride along a highly scrutinized cost-benefit decision matrix of rent seeking behavior off IBM's contracts with organizations that had already embraced those tools.
Billions of tokens have been burned in creating PPTs and Excel sheets, and emails discussing them, about why the Boundary Community Edition should be archived and abandoned.
I feel like I’m seeing a sleeper cell come to life. Your account is 5 years old with a handful of comments. You’ve been lying in wait for years only for today to be the day you get activated to eviscerate big corporate tech culture… and I love it. Billions of tokens writing reports, then another billion summarizing them, then another billion creating a followup, then another billion…
I recall talking to a HashiCorp rep about Boundary and the ability to find a place for it. Still seems kind of niche. I guess it never really found its place as a CE version.
No. Infisical is headed the same direction. Most of the secrets management and PAM bits that matter are not even in the open source repo, you need a paid license to get access.
Infisical was really amazing. It did precisely what I needed and any missing pieces would go in as a PR that would be quickly reviewed and merged. Then came talk about raising more funds, signing up more enterprise customers. Then came refactors to "improve the architecture" and "increase security". All the core pieces I depended on were moved behind an enterprise license.
Want Infisical to support Forgejo logins or runners via keys minted by step-ca? Enterprise license. Special secret cubbys or enclaves? Enterprise license.
Anything outside a toy usecase for a home user - Enterprise license.
A year ago I spent ~3 focused weeks ripping out every piece of Infisical in my infra and replacing it with a mix of Keycloak, Conjur and Age/SOPS. They are ugly, hard to work with, setup and debug - I'm not proud of this monster but I won't have to deal with what's going to happen to Infisical.
Infisical has every right to do this, I'm not complaining about that. I'm simply mentioning that we have seen this movie before, including the company in the title, so at this point, the tells are clear.
That might work for end user facing services but isn't the best fit for infrastructure. For infrastructure, you need something that people in the business understand well, are comfortable using, speaks the same language of operations so that vocabulary is well understood, and battle tested in production.
That's why Kubernetes has gained a lot of traction, not because 99% of companies need the scalability of Kubernetes, but it is a well understood concept, suite of services, and the vocabulary is very common now and well understood.
Nomad is a great fit for most compared to Kubernetes but failed to gain adoption because of these reasons.
An accurate title would be "IBM archives the IBM-HashiCorp Boundary Community Edition repo".
That "HashiCorp" in the title is a farce. The "Hashi" of the HashiCorp is far gone. IBM and "Corp" are redundant.
Community Editions don't make sense to the MBAs at IBM unless killing those off would create a global backlash (loss of goodwill for no positive revenue outcome is not good to an MBA. If loss of goodwill however, results in positive revenue, that's a good thing to an MBA. This is a well regarded, well studied and highly respected, revenue boosting MBA concept called "enshittification"). MBAs know that anything that doesn't directly translate into either viral growth or revenue next quarter is a waste of investor dollars. That makes anything else a risk to their own careers.
Boundary is past its viral growth stage, so the MBA needs to see revenue next quarter. Boundary fails the test there, so off with it!
HashiCorp was an innovative company founded by two engineers that built a lot of foundational, transformational infra products. Then, in a growth frenzy based off the traction of these tools they raised tremendous, vast sums of money and subsequently ~lost control of their vision~ hired MBAs, and the company went down the drain as well, or as MBAs call it - got acquired.
What remains of those tools are a BSL that focuses on maintenance operations and feature requests that strictly ride along a highly scrutinized cost-benefit decision matrix of rent seeking behavior off IBM's contracts with organizations that had already embraced those tools.
Billions of tokens have been burned in creating PPTs and Excel sheets, and emails discussing them, about why the Boundary Community Edition should be archived and abandoned.
I feel like I’m seeing a sleeper cell come to life. Your account is 5 years old with a handful of comments. You’ve been lying in wait for years only for today to be the day you get activated to eviscerate big corporate tech culture… and I love it. Billions of tokens writing reports, then another billion summarizing them, then another billion creating a followup, then another billion…
I recall talking to a HashiCorp rep about Boundary and the ability to find a place for it. Still seems kind of niche. I guess it never really found its place as a CE version.
hmm what is a good alternative to this ?
I used https://pomerium.com in the past for homelab. It builds around https://envoyproxy.io.
You can use Infisical for both secrets, certs, and PAM: https://infisical.com
No. Infisical is headed the same direction. Most of the secrets management and PAM bits that matter are not even in the open source repo, you need a paid license to get access.
Infisical was really amazing. It did precisely what I needed and any missing pieces would go in as a PR that would be quickly reviewed and merged. Then came talk about raising more funds, signing up more enterprise customers. Then came refactors to "improve the architecture" and "increase security". All the core pieces I depended on were moved behind an enterprise license.
Want Infisical to support Forgejo logins or runners via keys minted by step-ca? Enterprise license. Special secret cubbys or enclaves? Enterprise license.
Anything outside a toy usecase for a home user - Enterprise license.
A year ago I spent ~3 focused weeks ripping out every piece of Infisical in my infra and replacing it with a mix of Keycloak, Conjur and Age/SOPS. They are ugly, hard to work with, setup and debug - I'm not proud of this monster but I won't have to deal with what's going to happen to Infisical.
Infisical has every right to do this, I'm not complaining about that. I'm simply mentioning that we have seen this movie before, including the company in the title, so at this point, the tells are clear.
Any reason you chose conjur instead of openbao? (the guaranteed open source/no restriction fork of hashicorp vault)
IBM archives the Boundary open source repo
and the license makes it radioactive -unusable.
Just train your llm on it and create virtually the same thing it’s like a copyright get out of jail free card.
That might work for end user facing services but isn't the best fit for infrastructure. For infrastructure, you need something that people in the business understand well, are comfortable using, speaks the same language of operations so that vocabulary is well understood, and battle tested in production.
That's why Kubernetes has gained a lot of traction, not because 99% of companies need the scalability of Kubernetes, but it is a well understood concept, suite of services, and the vocabulary is very common now and well understood.
Nomad is a great fit for most compared to Kubernetes but failed to gain adoption because of these reasons.
It was using the Mozilla Public license until August 2023 so versions before then might be usable. Some parts are still under MPL
https://github.com/hashicorp/boundary/commit/29da0bcb923236b...