Never having to deal directly with a version control system, or even a static site generator, on your own personal computer is up to you. The online IDE is designed and ready for your browser-only workflow whenever you are. In a pinch, even the mobile (browser) interface will get it done, instantly.
Multi-gigabyte source trees are simply no match for it — full site build durations are measured in seconds, not hours and minutes. The system easily sustains over
500 MB/s while processing your sources for full site builds, but given its unique, patent-pending dependency management functionality, it almost never needs to resort to them:
It only builds what you need built.™
This technology was first invented in late 2011 by members of the Apache Infrastructure Team. The phrase “JAM Stack“ was coined years later, by a group unaware of our prior art in this space. But the public record is crystal clear, and this new offering by SunStar Systems is testimony to that original legacy and its true open source origins.
The system we created for The ASF has been running on autopilot for the past 6 years. The only issues we had with it from a stability standpoint revolved around ZFS usage on FreeBSD 9.1, which was a work in progress back then. The fact is that this CMS/IDE is just the CI / CD coordination hub for power users on CDN services (or similar), and those CDN’s are where the action is for SLA performance on the JAM Stack. If our CMS system goes down, your ability to modify your live site will be impacted, but the last-mile delivery to your actual site’s consumers is unimpacted thereby. If The ASF’s history with this system is a guide, the likelihood of an unplanned downtime event is relatively remote, especially when compared to the version control system which underpins it.
Want to give it a quick spin to see how fast and easy it is?
This website is self-hosted; all those hot-pink pencil squares at the top right are live. Feel free to poke around and experiment with anything you see! The only thing you can’t do is commit and watch your changes build and deploy; that’s restricted to our personnel for obvious reasons.
Our global Point of Presence (POP) network ensures low-latency connections to both the CMS and the end-delivery websites we host. We are currently on four continents and have a bicoastal presence within the US. And we’re expanding rapidly- next up East Asia!
|CI / CD||Y||Y||N|
|FedRAMP Infrastructure + FIPS compliant TLS||Y||N||N|
|Browser-Only Workflow (Mobile Friendly)||Y||N||N|
|Unlimited Branch Builds||Y||Y||Y (for public repos)|
|Scheduled Builds of Dynamic Content Only||Y||N||N|
|Easy 3rd Party Contributor Workflow||Y||Y||Y|
|No Per-Preview Billing||Y||Enterprise Only ($$)||Y|
|No Per-Build Billing||Y||Enterprise Only ($$)||Y (for public repos)|
|No Per-Seat Billing||Enterprise Only||N||Y (for public repos)|
|SSI / CGI / mod_perl registry Support||Y||N||N|
|Fine-Grained ACL on Source Repo||Enterprise Only||N||N|
|+500 MB/s Full Site Builds||Y||?||N|
|8-64 Way Concurrent Builds||Y||Enterprise Only ($$)||N|
|TCO < $5000/yr for Enterprise||Y||N||?|
|99.99% SLA for Enterprise||Coming in Q4||Y||N|
|Serverless Integration||Coming Soon!||Y||N|
|Choice of Programming Language||Coming Soon!||N||N|
Informed consumers should always ask which components of the JAM Stack are covered by SLA, or get written assurance that your SLA provides blanket coverage for the entire stack, including the VCS, build infra, and CMS / IDE portions. You might be surprised by the answers: even GitHub has serious multi-hour outages from time to time, on which many competitors base their entire offerings. Such services with nominal 99.99% SLA’s, which is less than an hour a year of total downtime, clearly have some explaining to do, if you are under the impression your full service stack is under SLA. It’s not, we’re just being up front with you about it.
What we’re offering instead, for 2020, is a best effort to keep our Subversion service online 24x7x365, with primary focus on Continental US business hours, and with a competitive target of having our SLI‘s read favorably when compared to the rest of the market (including GitHub itself, whose SLA’s revolve around response time targets, not the actual duration of downtime incidents). SLA’s are an important ingredient, so Enterprise customers will enjoy our comprehensive 99.99%, which covers everything but the Subversion service itself. Nobody offers blanket availability SLA contracts for their version control service infrastructure; and you should be wary of doing business with someone who does in today’s marketplace.
It does all this by either interfacing with an svnpubsub-enabled, customer hosted Subversion service, or by using one of ours. It automatically builds each committed change using a lightning fast
Perl 7 build system based on the Hollywood Principle: it calls out to customer-supplied Perl modules to do whatever the customer needs done (don’t worry, there are basic layout templates available that will do all the heavy lifting for you). It also goes the extra mile and instantly distributes the change-sets for the build, atomically, to your live website servers / CDN providers.
REST compliance has its benefits to CMS users. They’ll like the fact that their browser’s Back button doesn’t break the service; in fact it’s an essential component of the Rollback functionality. In an emergency site breakage situation, going back to the bad Commit submission’s result page will provide them with the ability to reverse merge that errant applied change — at the click of a button — and rebuild the site back to a known good state.
We currently provide a basic Django based template system derived from a customized
Dotiac::DTL module from CPAN. Support for a Python 3 build system is coming in 2020 Q4 (other languages will follow in 2021). The Apache 2.0 Licensed CMS build system, together with a large collection of supporting modules to take advantage of within that tree, is available on GitHub.