Database-Based Operating System 'DBOS' Does Things Linux Can't (nextplatform.com) 104
Databricks CTO Matei Zaharia "said that Databricks had to keep track of scheduling a million things," remembers adjunct MIT professor Michael Stonebraker. " He said that this can't be done with traditional operating system scheduling, and so this was done out of a Postgres database. And then he started to whine that Postgres was too slow, and I told him we can do better than that...."
This resulted in DBOS — short for "database operating system" — which they teamed up to build with teams Stanford and MIT, according to The Next Platform: They founded a company to commercialize the idea in April 2023 and secured $8.5 million initial seed funding to start building the real DBOS. Engine Ventures and Construct Capital led the funding, along with Sinewave and GutBrain Ventures...
"The state that the operating system has to keep track of — memory, files, messages, and so on — is approximately linear to the resources you have got," says Stonebraker. "So without me saying another word, keeping track of operating system state is a database problem not addressed by current operating system schedulers. Moreover, OLTP [Online Transaction Processing] database performance has gone up dramatically, and that is why we thought instead of running the database system in user space on top of the operating system, why don't we invert our thinking 180 degrees and run the operating system on top of the database, with all of the operating services are coded in SQL...?"
For now, DBOS can give the same kind of performance as that full blown Linux operating system, and thanks to the distributed database underpinnings of its kernel, it can do things that a Linux kernel just cannot do... One is provide reliable execution, which means that if a program running atop DBOS is ever interrupted, it starts where it left off and does not have to redo its work from some arbitrary earlier point and does not crash and have to start from the beginning. And because every little bit of the state of the operating system — and therefore the applications that run atop it — is preserved, you can go backwards in time in the system and restart the operating system if it experiences some sort of anomaly, such as a bad piece of application software running or a hack attack. You can use this "time travel" feature, as Stonebraker calls it, to reproduce what are called heisenbugs — ones that are very hard to reproduce precisely because there is no shared state in the distributed Linux and Kubernetes environment and that are increasingly prevalent in a world of microservices.
The other benefit of the DBOS is that it presents a smaller attack surface for hackers, which boosts security, and that you analyze the metrics of the operating system in place since they are already in a NoSQL database that can be queried rather than aggregating a bunch of log files from up and down the software stack to try to figure out what is going on...
There is also a custom tier for DBOS, which we presume costs money, that can use other databases and datastores for user application data, stores more than three days of log data, can have multiple users per account, that adds email and Slack support with DBOS techies, and that is available on other clouds as well as AWS.
The operating system kernel/scheduler "is itself largely a database," with services written in TypeScript, according to the article. The first iteration used the FoundationDB distributed key-value store for its scheduling core (open sourced by Apple in 2018), according to the article — "a blazingly fast NoSQL database... Stonebraker says there is no reason to believe that DBOS can't scale across 1 million cores or more and support Java, Python, and other application languages as they are needed by customers..."
And the article speculates they could take things even further. "There is no reason why DBOS cannot complete the circle and not only have a database as an operating system kernel, but also have a relational database as the file system for applications."
This resulted in DBOS — short for "database operating system" — which they teamed up to build with teams Stanford and MIT, according to The Next Platform: They founded a company to commercialize the idea in April 2023 and secured $8.5 million initial seed funding to start building the real DBOS. Engine Ventures and Construct Capital led the funding, along with Sinewave and GutBrain Ventures...
"The state that the operating system has to keep track of — memory, files, messages, and so on — is approximately linear to the resources you have got," says Stonebraker. "So without me saying another word, keeping track of operating system state is a database problem not addressed by current operating system schedulers. Moreover, OLTP [Online Transaction Processing] database performance has gone up dramatically, and that is why we thought instead of running the database system in user space on top of the operating system, why don't we invert our thinking 180 degrees and run the operating system on top of the database, with all of the operating services are coded in SQL...?"
For now, DBOS can give the same kind of performance as that full blown Linux operating system, and thanks to the distributed database underpinnings of its kernel, it can do things that a Linux kernel just cannot do... One is provide reliable execution, which means that if a program running atop DBOS is ever interrupted, it starts where it left off and does not have to redo its work from some arbitrary earlier point and does not crash and have to start from the beginning. And because every little bit of the state of the operating system — and therefore the applications that run atop it — is preserved, you can go backwards in time in the system and restart the operating system if it experiences some sort of anomaly, such as a bad piece of application software running or a hack attack. You can use this "time travel" feature, as Stonebraker calls it, to reproduce what are called heisenbugs — ones that are very hard to reproduce precisely because there is no shared state in the distributed Linux and Kubernetes environment and that are increasingly prevalent in a world of microservices.
The other benefit of the DBOS is that it presents a smaller attack surface for hackers, which boosts security, and that you analyze the metrics of the operating system in place since they are already in a NoSQL database that can be queried rather than aggregating a bunch of log files from up and down the software stack to try to figure out what is going on...
There is also a custom tier for DBOS, which we presume costs money, that can use other databases and datastores for user application data, stores more than three days of log data, can have multiple users per account, that adds email and Slack support with DBOS techies, and that is available on other clouds as well as AWS.
The operating system kernel/scheduler "is itself largely a database," with services written in TypeScript, according to the article. The first iteration used the FoundationDB distributed key-value store for its scheduling core (open sourced by Apple in 2018), according to the article — "a blazingly fast NoSQL database... Stonebraker says there is no reason to believe that DBOS can't scale across 1 million cores or more and support Java, Python, and other application languages as they are needed by customers..."
And the article speculates they could take things even further. "There is no reason why DBOS cannot complete the circle and not only have a database as an operating system kernel, but also have a relational database as the file system for applications."
Time travel OS (Score:5, Insightful)
Interesting ideas but there are cases where you still won't be able to "time travel" your system state to repeat some past event to hunt bugs or whatever. Off the time of my head is anything that cares what the current time is. Another is anything that relies on a RNG to produce results. And I don't have a specific example but gut feeling tells me there's a potential security issue if you can roll back time to a prior state and try something repeatedly involving a password or encrypted data or something... not sure if this is true but feels uncomfortable in a security context.
Re:Time travel OS (Score:4, Interesting)
I'd assume theres some sort of software defined state serialization/deserialization, similar to the way Apple's stuff works where if the system wants to freeze an app, it tells the app to store its shit however it needs to, then on rehydration, it does the inverse. Thus its up to the programmer to figure out that stuff. If something depends on externalities (time, hardware states, etc) then the programmer needs to fine a sane solution to it.
This is kind of a solved problem.
Re:Time travel OS (Score:4, Informative)
Not really, it just kicks the problem out a level. There's a similar mechanism on Android. It's probably in the top 5 things new developers don't understand and need to be taught in detail. It's doable, if the devs writing the program understand and correctly use the system, but it's not free or anywhere near a solved problem.
Re: (Score:2)
You could simulate the previous date/time. Perhaps by changing the system clock or have the API apply an offset, Shirley.
Re: (Score:3)
Re: (Score:3)
This was the principle behind OS/400, the original operating system of the IBM AS/400.
Re: (Score:3)
Re: (Score:3)
Sure. But they are really just re-implementing old ideas and claim they are new to get investors. It is well known that in databases or any scheduling systems really, repeatability requires limits on what you can do. You already found two of those by just thinking about it for a minute. There are some more. You also get performance bottlenecks when you have a specific scheduling order.
The whole thing has been tried before and it remained something for a niche. Of course they do not tell that to the "investo
Re: (Score:2)
All true - but if they were to commercialise and actually make it good, then having a cluster of machines that wake up at the right times, run jobs and report results could be a handy feature of many application stacks. Of course, we all do the same already, except our current solutions all suffer from problems if things crash, or if inputs are out of range or whatever. We've glossed over and 'normalised' those problems for so long we probably don't even know we have a problem that could do with a decent so
Re:Time travel OS (Score:4, Interesting)
Four things that seem to be missing:
1. A free open source license
2. Visible source so as to improve the code by others
3. SQL extensions to support OS needs including high availability/hardware tolerance, synchronization with the pub/sub model context, data bus rail integrity, and networking shims for flow flexibility and audit.
4. Comparison (in the cited article) with optimized Linux platforms, rather than poising it simply against Amazon.
It all smells like a VC hustle.
Re: (Score:2)
A lot of companies won't (and probably shouldn't) care about items 1 and 2, and won't care that much about 4.
Re: (Score:3)
Opaque and closed source is state of the art 1995.
A lot of *developers* won't trust a platform they can't see, only react to. And as licensing and monetizing is an important characteristic of business relationships today, they're quite interested in how they'll interact with a database blob.
Re: (Score:2)
Are you sure about that? The last large-scale closed-source disaster (VmWare) is still in progress and many, many organizations scramble to deal with it. There are enough others that many organizations will have noticed.
Re: (Score:2)
Re: (Score:2)
I'm guessing it's another one, an RTOS: https://www.theregister.com/20... [theregister.com]
They invented (Score:5, Interesting)
OS/400 all over again. https://en.wikipedia.org/wiki/... [wikipedia.org]
Re: (Score:2)
That's the first thing that I thought of.
They've borrowed some of the concepts from OS400/IBM i.
Though probably not the sophistication of the dual-abstraction concept.
Ghost of Dick Pick (Score:5, Informative)
The Pick Operating System [wikipedia.org] (developed by Dick Pick) was a database running straight on the metal. In some cases, the host system had customised microcode to run the Pick system bytecode natively.
Re: (Score:3)
The Pick Operating System [wikipedia.org] (developed by Dick Pick)
Does Mr Dick Pick send unsolicited images of his OS?
Re:Ghost of Dick Pick (Score:4, Informative)
Sadly he died in 1994 at the age of 56 (complications following a stroke). He didn't live to see camera phones usher in the dick pick revolution.
Re: (Score:2)
Re: (Score:2)
Thank you! I was going to post something along these lines but wanted to see if anyone else did too. There was a time nearly every single automobile and motorcycle dealership ran Pick/OS to maintain their inventory and ordering systems! I don't know if this is still true.
Re: (Score:1)
That must be a Wikipedia prankster. In the web age it's just too hard to accept that as a legitimate name. I'm more likely to believe Jesus Christ is really Al Bundy.
Re: (Score:2)
Re: (Score:2)
TIMS nonsense? TIMI is great! It allowed multiple CPU architecture changes without breaking any software.
Pick Operating System redux (Score:5, Insightful)
The Pick System, aka Reality, implemented this idea back in 1965. I suspect DBOS will end up on the same route Pick traveled.
Re: (Score:2)
Next up someone is going to "discover" the thin client / fat server model.
Re: (Score:2)
That's scheduled for 2025, with the fat server located in the cloud, so they can properly bill the thin clients.
Re: (Score:2)
They'll be going broke, then. Have you seen most IT guys?
Re: (Score:2)
I read the comments to see if anyone would mention Pick.
I'm off to have PTSD, now.
Re: (Score:3)
Long and interesting history. https://en.wikipedia.org/wiki/... [wikipedia.org]
Re:Pick Operating System redux (Score:5, Funny)
Re: (Score:3)
It's highly specialized, but it has its uses. I love Pick. Even emulated as it is today, it's still fast as hell and very reliable
Re:Pick Operating System redux (Score:5, Interesting)
Re: (Score:2)
Well, "those that do not understand existing technology will re-implement it, poorly" or something like that.
"Y does something X can't" (Score:2)
Doesn't mean Y is better than - or even as good as - X for most tasks that X is currently used for.
A seaplane can do something most airplanes can't: Land on and take off from water. That doesn't make it the best option for most air travel. But for some use cases, it's better than a conventional, "land and takeoff on wheels" airplane.
I expect there will be some cases where this database-based system is better than a traditional Linux kernel-based system, but I'm not seeing a compelling reason to consider
Re: (Score:1)
I expect that someone will read this article and the key features will be bolted onto Linux within a year, just because Linux developers are serial contrarians.
Re: (Score:2)
Indeed. Linux also cannot boil an egg. This "Linux" thing really seems to be a joke, no?
Re: (Score:1)
Linux can't boil an egg? Don't be so [cnx-software.com] sure [allrecipes.com].
Soft-boiled eggs, Penguin-style!
Ok (Score:2)
Sounds like a fine project.
Why the rude gesture at Linux?
not Ok (Score:2)
either they dont really understand all the intricacies of a OS to even support python or some other useful interpreter.... they will need more than a scheduler they need to manage IO and have drivers for that... for example are they going to wrote their own TCP/IP stack ?
I dont think so.... these basic questions should have been asked by the article
the IO subsystem and compatibility are its major problems... Linux had exactly the same issues until people started contributing and the GPL forces everyone to s
Re: (Score:3)
Why the rude gesture at Linux?
Because you need to take a piss on the competition if you want your snake-oi^W^Wexcellent new invention to be interesting to investors because "disruption" is a key word for spontaneous investor-orgas^Winterest.
Based on NoSQL? No ACID compliance? (Score:3)
Stopped reading at "NoSQL". For a second, I thought this might solve a problem with existing filesystems, which is giving assurance to a remote program that a change in a file is now stored on media. In most filesystems, you can't guarentee this, because it would slow things down, and a lot of stuff will just cache stuff in RAM, return "yah, dude, its on disk... trust me". There is no real way to ensure that a file based transaction has made it or not. If this OS were ACID compliant, it would be nice, just ensuring data was guaranteed to be on the disk, but with NoSQL, I have no assurance of anything.
It would be nice to see an OS that can solve this, or even a filesystem type, but this seems something like IBM OS/400 or PICK, and other than a few environments, those are pretty much bit players.
Re: (Score:2)
The intro claimed they tried PostgreSQL but it was too slow. Maybe they could tweak the PostgreSQL engine or schema to improve performance for OS needs.
If they had to nuke ACID to get satisfactory performance, could that be made a switch so that the shop can choose between ACID or speed? (That didn't sound right :-) I hope HR doesn't see this.)
It would be interesting to see the tradeoff analysis.
Re: (Score:2)
Mr. Tabilizer, please report to the Personnel Manager's office at 8:00am tomorrow (Monday, March 18) for a mandatory meeting. We ask that you bring your building keys.
Re: (Score:1)
I have Monkey Pox, I can't.
Re: (Score:1)
That's the bit that puzzles me. First they say PostgreSQL is too slow. Then later they say OLTP performance is fast. Then later they say Typescript services etc. Is Typescript stuff faster than PostgreSQL?
Maybe someone forgot to create some indexes in their PostgreSQL DB or: https://www.enterprisedb.com/b... [enterprisedb.com]
Re: (Score:1)
How about release it with PostgreSQL as open source, and let PostgreSQL experts experiment and tune it.
Re: Based on NoSQL? No ACID compliance? (Score:2)
How do you think postgres does it? It is solved you use fd* calls to adjust write caching per descriptor.
Re: (Score:2)
Re: (Score:2)
In most filesystems, you can't guarentee this, because it would slow things down, and a lot of stuff will just cache stuff in RAM, return "yah, dude, its on disk... trust me". There is no real way to ensure that a file based transaction has made it or not.
They are called synchronous writes, and every modern filesystem has a synchronous mode which can be accessed EITHER by applications requesting it, or through mount options which force it for ALL writes. The writes do not return until they are actually written. Yes, it slows things down, which is why you don't ask for it unless you actually need it.
Re: (Score:3)
IIRC, the Pick OS used a B+Tree to handle thse problems. I'm not sure if it was ACID compliant, but it was quite good.
This is just IBM i or as it was known AS/400 (Score:5, Interesting)
This is literally what AS/400 or as it is called now, IBM i, does. File I/O with read/write/update op codes are one-to-one select, insert, update SQL commands. In fact, the SQL precompiler literally converts exec sql statements into a HLL that just does file I/O. It's just all database under the hood. The way they talk about how scheduling works is really close to what IBM's OS already does.
Stonebraker says there is no reason to believe that DBOS can't scale across 1 million cores or more and support Java, Python, and other application languages as they are needed by customers...
You know what's crazy is that most people believe that IBM i just does the old COBOL/RPG stuff. But modern IBM i has a fairly modern C++ compiler and includes python, node.js, PHP, and Java among other things. And since the RedHat purchase, there's a bigger move to Linux like containers with Merlin [ibm.com]. Heck they even have VS Code running in a container, so that you can write RPGLE (and modern RPG looks a lot like C now) in VS Code via your web browser.
So I'm not really sure what DBOS is really offering that IBM i doesn't already do circles around. Yes, it's complete lock in with IBM's offering, nobody will debate anyone else on that. But IBM's POWER machines and their OS are specifically written for each other. Modern RPGLE programs can easily smoke a lot of the Wintel stuff, but mostly because the hardware, the compiler, the entire stack from start to finish are all finely tuned to work with each other. And if you have an IBM contract, the second any of your disks look like they're about to fail, or your controller cards are acting funny, IBM is there within a couple of days to replace the defective part no questions asked. When they say it is turnkey, they mean just that.
So yeah, IBM i isn't exactly the most sexy bit of technology out there and boy can it be a bitch to get going. But holy shit, once it is running it will never stop so long as electricity is getting to the box. And does the whole "I a need a database and I need to methods for accessing that data" very nicely because that is exactly what it was written for. It's absolutely not ideal to be running a whole website off it, it's a pretty poor front end OS. But if you have a database and want a RESTful API to access it, you'd be hard pressed to find a better choice. There are pretty good options and some of them don't require you to sign your soul over to IBM in exchange of "you take care of your hardware yourself". But IBM i absolutely does everything that DBOS indicated it does and likely does it millions folds better. I've not used DBOS, so who knows?
But I read all of that and instantly though, they're just reinventing IBM i.
Re: (Score:2)
> Modern RPGLE programs can easily smoke a lot of the Wintel stuff, but mostly because the hardware, the compiler, the entire stack from start to finish are all finely tuned to work with each other.
Integration can indeed result in productivity. When programming languages and IDE's used to be closely aligned, it was much less time and code to code the same non-large CRUD app. The web gave us a lot of flexibility, but at the cost of roughly 3x the coding labor and code. Web stacks are focking bloated layer
Re: (Score:3)
Re: (Score:2)
I always knew that Dungeons and Dragons had arcane rules and cryptic game play mechanics - barely distinguishable from compute code!
Re: (Score:2)
The problem is that IBM is really bad at pushing this technology for some reason; I've worked at places that run OS/400 and the only thing the managers want is to get rid of it all, while at the same time keep it running forever.
So I've seen people build tons of typescript microservices on top of DB2 tables that are also used by ancient, almost untouched RPG/Cobol systems, but these microservices run on AWS or a local kubernetes, but never really utilizing these wonderful technologies you mention.
And in the
Re: (Score:2)
I can't really speak for anyone else, but for us, it was the combination of the expense and the vendor lock-in. If I remember correctly, the straw that broke the camel's back was when we tried to connect a non-IBM SCSI tape
Re: (Score:2)
> I've not used DBOS, so who knows?
Perhaps the BDOS developers do.
Tablegasm! (Score:2)
This makes we table-izers salivate! IBM's AS400 alleged did such, but it mostly got ignored because it was proprietary. Hierarchical file systems are too damned limiting. I agree we should keep the path idea, but also allow searching, indexing, and grouping on file/folder attributes of our own choosing*.
Hopefully it will re-spark table-oriented-programming, which had hints of taking off under CASE tools before OOP stole the show. Even if it turns out a dud, it couldn't be worse than all the other dud fads t
Re: (Score:1)
Minor corrections:
1. ...all the other dud fads that have come along.
2. ...you just get Nulls, unless you switch the table to "static mode".
Where are OSS GrammarGPT browser pluggins, dammit!
"operating services are coded in SQL" (Score:2)
What does that even mean? To open a file you write SELECT contents from filetable where name="retards.txt"
FFS. Are they fucking stupid? Or maybe they fucking genius. How can they come up with something so asinine, and then manage to dupe investors with that madness?
Re: (Score:2)
How can they come up with something so asinine, and then manage to dupe investors with that madness?
I take it you've only recently arrived here on Earth...
Re: (Score:2)
If you really need a database backend, and you are ready to sell your soul to IBM, OS/400 (on the hardware platform AS/400) respective IBM i for Business running on IBM System i is quite nifty.
Re: (Score:2)
This "description" is not targeted at people with some understanding of technology. Investors frequently have no clue as to the nature of what they invest in. It is a form of high-stakes gambling essentially. People with some actual understanding can often see that these things are not new, will go nowhere or are just really bad ideas that sound good when you have limited understanding.
If you look at the mountains of money that are currently made with moronic, clueless and hallucinating "AI", you can see ho
Re: (Score:2)
1. Slide deck: "The market for product $XXX and we believe we can capture YY%." (based on merely the fact that they exist as a product) ........
2. Finance guys write checks.
3. Market share never achieved or profit margins are not nearly what was projected or both.
4. IPO and offload massive debt to second round of suckers.
5. Pivot to something else.
Re: (Score:2)
Actually it sounds like a very good idea. Just for a niche usage case, and not really new (but perhaps there are new optimizations, or specialized hardware that can be optimized for).
Re: (Score:2)
A database view is a subset of a database that is based on a query that runs on one or more database tables. Database views are virtual tables that present data from one or more tables in a absolutely consistent
Lambda (Score:2)
It's hard to read this rambling article, but apparently this is supposed to compete with Lambda? Makes sense, the biggest struggle with running containerized event-driven workloads is just dealing with that bloated operating system underneath. This is kind of interesting. If it works, I'm sure it will just be bought by Amazon or Google or Microsoft.
PR doesn't pass the sniff test (Score:2)
And the resuming of operations is different from any other operating system with preemptive multitasking? Which program that was preempted has to redo its work from previous states and most current operating systems? (Let's skip low-level locks implemente
Oversimplification (Score:3)
Not all the operations that an OS makes are IO operations. I take this article to be some kind of sales pitch for a database software company who is trying to advertise how fast their database is, but I fail to see what database speed has anything to do with being an operating system.
If this is about being some sort of ACID file system, i suppose it has academic interest, but other than being a marketing ploy, what is the practical use of this over traditional file systems?
Re: (Score:2)
If this is about being some sort of ACID file system, i suppose it has academic interest, but other than being a marketing ploy, what is the practical use of this over traditional file systems?
Database file systems are not exactly new, as are hybrid products (filesystems with database-like interfaces, databases with file storage, etc.)
None of the previous ones have gone anywhere. I don't see how this one would be much different, and the fact that they use Typescript yells "get rich quick start-up scheme".
Re: (Score:2)
and the fact that they use Typescript yells "get rich quick start-up scheme"
It does. As does the invalid comparison to Linux (apples and oranges).
Re: (Score:2)
Re: (Score:2)
Take a look at the Pick operating system or IBM OS/400 resp. IBM i to look how a database can be an operating system all on its own.
I'm not familiar with Pick, but I have a little bit of familiarity with OS/400, so I know that is nonsense. It is a LOT more than just DB/2. A database alone can technically be an OS, but only a bad one with a whole lot of missing features and really lousy performance. They are touting their performance... by saying it's as good as Linux! Well, uh wow, guess what? We have another operating system that's as good as Linux, we call it... Linux.
Re: (Score:2)
If you compare Linux to any of the three systems, you will find that they are much bett
Re: (Score:2)
First, operating system does not mean operating system with a shell for desktop work.
No one said it did.
OS/400 was not meant to be a machine to install a word processor on or browse the World Wide Web.
No one said it was.
An operating system at first is a program (or a set of programs) to administer all resources of a computer and share them between the applicartions.
An operating system is the collection of the software that lets you do that (usually a kernel, though in a microkernel-based system, it might be a kernel and some servers) plus the software necessary for day to day maintenance, plus the software that allows you to interface with the system. Without all of those things it is not a complete OS. You just described part of an OS.
The way you act with the operating system is the shell.
That's one way you can interact with an operating system, yes, but it is hardly the only way. Fo
Re: (Score:2)
Indeed. The claim "better than Linux" is nonsense as they are comparing two different things. What they have (or may have) is a database system with specific optimizations. It is not unheard of for database systems to run on bare metal with no OS in between.
Re: (Score:2)
Also, TFA states:
The first iteration of DBOS runs on Amazon Web Services and uses the Firecracker microVM service, itself a stripped down KVM hypervisor running on a stripped down Linux microkernel, to create the user space for FoundationDB to run within. So technically, there is still Linux underneath DBOS . But nothing like the full blown Red Hat Enterprise Linux or SUSE Linux Enterprise Server that companies deploy or the homegrown, full-blown Linuxes that the hyperscalers and cloud builders have created for their own use.
Re: (Score:2)
Yes, no, maybe (Score:2)
It's really a reimplementation of a couple of older OS'. But so is Linux. Linux did Unix better than many commercial Unices, so won. It'll be 10-15 years before we'll discover if this new OS is a better AS/400. I think we should allow for that possibility.
But it doesn't sound like the developers have a good understanding of what has been tried or what is needed for the OS to be useful. So I'm not holding my breath. Still, that can change.
From what I'm seeing, this OS might be interesting for running virtual
Re: (Score:2)
No, Linux didn't reimplement UNIX in a better way. Linux dominate over UNIX for two reasons:
1) The suit against AT&T which prevented them from pushing UNIX
and
2) The GPL vs the BSD licenses. The GPL encouraged software to be modifiable and freely available.
The BSD license is better of system standards, but for pushing particular software programs, the GPL is better, because it discourages hiding the changes you made to the code. Which license is better depends on how you want to use it, but the GPL w
Apples and Oranges (Score:4, Insightful)
This is not an OS, this is a scheduling system. Comparisons to Linux and any other OS are just nonsense.
Re: (Score:2)
Re: (Score:2)
Written in SQL (Score:4, Funny)
Re: (Score:3)
Writing an entire OS in SQL sounds like a special kind of torture.
The Unix equivalent would be to write the entire OS as a series of shell scripts. Nobody does that, despite the fact that it's completely viable. The everything's a file approach (with the existence of the proc and sys filesystems) means that shell scripts can do everything your userland can do now... only more slowly.
In order to get performance better than other systems, if that is even possible, they will have to abandon that approach and rewrite parts of the OS in another system, presumably whatever thei
Just like others....... (Score:1)
fits into RAM versus big data (Score:2)
The kernel can keep track of open files because the descriptors all have to fit into RAM. You don't need a database for that, and you are not likely to need to search for multiple keys or do joins. If a person is going to be an expery in databases, they need to understand there is a purpose and place for simple data structures.
Reminds me of (Score:2)
BeOS File System (BFS) (Score:2)
Heisenberg Compensator (Score:1)
Welcome to 1965 (Score:1)
Re: (Score:2)
He means my mom, specifically. This is a directed taunt at me, which he thinks will jostle free some suppressed memories and trigger an emotional breakdown. What he thinks I don't know is that that's not even really my mom. What I know that he doesn't though, is that not only is that not my mom, it's not even the person who he thinks is pretending to be her. Going from just the memories I actually have, she's at least the 3rd consecutive Russian spy in a row to have killed the previous one and stolen custod
Re: (Score:1)
Ah, I knew I should've waited for Rocky and Bullwinkle to answer my email about the information Boris and Natasha mistakenly left on their front porch before I posted here. Frostbite Falls's very limited internet [wikipedia.org] must be down again.