A pretty good article and I am looking at Yugabyte more seriously (again).
But I still have questions:
1. Is CockroachDB actually faster (generally) than Yugabyte? Their (own) benchmarks seem to show it is.
2. If true, why? Did writing their own pgsql-compat interpreter really make that much of a difference?
3. This is the big one for me: why does Yugabyte have such large percentages of code in its Github repo in different languages, especially Java, JS and TS? I expect to see a small amount of scripting languages like Python for building or testing or whatever, but it appears that around 30% of the code is actually Java, for a server that was built on top of pure Postgresql. (I'm not a fan of Java in general, except for Elixir.)
Other than those questions, Yugabyte looks like a pretty good migration target and their licensing approach is refreshing.
Writing a database from scratch is not easy. YugabyteDB uses some PostgreSQL, Kudu, and RocksDB code that has been heavily optimized before. Those are good codebases, and only some parts need to be enhanced to make them distributed.
2. Their Go version of RocksDB, Peeble, seems less efficient. They did it for a good reason. They didn't have the C++ skills to enhance RocksDB itself.
3. The repo holds more than the database.
C: is the SQL layer, based on PostgreSQL
C++: the transactional distributed storage, heavily modified Kudu and RocksDB
Java: some regression tests, the managed service automation, sample applications
TS: the Graphical User Interface
Python: some tooling to build the releases, some support tools
1. Nope, I don't think that CRDB is generally faster than Yugabyte. Vendors can't say otherwise in their own benchmarks :)
3. That code must be the tooling around the db and various smart drivers (that extend default drivers of Postgres for various programming languages). Overall, the core database engine of Yugabyte is written in C.
> That reveals a startling statistic that 80% of the market is controlled by only 12,000 users. That's 433 units per user on average.
This conflates the market for this type of software with the larger housing market. Presumably, this software is only used by some large (probably mostly hedge-fund or REIT funds) apartment complexes. So, yeah, 80% of that market, but that's only a tiny amount of the overall market.
The DOJ chose that number (almost certainly out of a hat, because there's no way to actually identify all of them) of 80% in order to make it seem like it's a monopoly, but proving at trial that RealPage or its clients controls prices over more than a tiny fraction of the market will be incredibly difficult (actually impossible, because it's obviously not true).
However, given the DOJ's limitless funding to pursue this, clearly it's not after actually winning at trial; it's after a consent decree.
This is not appropriate for the DOJ to be taking up. They have nearly unlimited resources and have far bigger fish to fry.
The appropriate enforcement agencies are the States' Attorneys General. If they feel that their constituents are being harmed, they can take it up with the courts.
There is certainly zero evidence that RealPage has any sort of monopoly. The DOJ would have named that evidence in their press release if they had any evidence at all that RealPage possessed any sort of monopoly power. This isn't the National Assoc of Realtors (which obviously does actually have strong monopoly power), and where is the DOJ when the NAR lawsuits have been going on for the last several years? So, the Sherman Antitrust Act claims are tenuous at best.
As it stands, this seems to be a ham-handed attempt to seek a consent decree for clumsy price-fixing, setting maximum rent prices, nationwide, without any appropriate legislation, as an end-run around Congress. And that will actually harm consumers.
>> this seem to be a ham-handed attempt to seek a consent decree to create minimum rent prices, nationwide, without any appropriate legislation. And that will actually harm consumers.
Where does this contention come from? What makes it “seem” like this is the goal of the DOJ to you?
The DOJ has not a prayer of proving that RealPage (or its clients or whatever) controls more than a tiny minority of the housing market, because it just doesn't. This is a tiny tiny fraction of all housing nationwide.
So, since it can't do that and actually win this case on that basis, but on the other hand has virtually unlimited resources to string this out indefinitely, it's after something else. So, what do you think that could be? My guess is that this is an attempt at forcing these complexes into a price-setting consent decree.
(by the way, thanks for actually responding with a criticism. A lot of time people just signal they disagree but they can't be bothered or can't articulate a real response. I do appreciate the substantive argument even if we might disagree)
How many Gates are there in a typical Harvard grad cohort? Are 5% of Harvard grads a Gates?
If we're looking at lightning strikes, it seems like we should look at some of the most successful tech CEO's in history, in no particular order.
Gates & Zuck - Harvard (Microsoft and Facebook; both dropped out by end of second year or earlier)
Larry Ellison (Oracle, third richest man in the world in 2000's and currently 9th richest) - dropped out of Univ Illinois and Chicago
Larry Page, Sergein Brin (Google, dropped out of Stanford)
Elon Musk (Tesla/SpaceX/X/PayPal/etc, richest man in world, attended Queen's and then transferred to ivy league U Penn, one of the very few that completed a degree, BS in both Physics and Econ)
Jeff Bezos (Amazon, second richest in world, earned BSE in EE/CS from Princeton, summa cum laude)
Steve Jobs (Apple, dropout Reed College)
Of course, these numbers have such a high p-value that really it amounts to nothing more than a guess or, at best, a hypothesis, and I get that you're just suggesting that those who both get admitted to Harvard and actually complete a full education there have much higher than average rates of being successful, but it seems that the extreme ends of such a graph end up being counter-intuitive, because:
From a very quick review of an admittedly hand-picked assortment of the very small number of "most" successful tech CEO's (whatever that might mean), we can identify a few common characteristics:
1. most did attend a university of varying quality.
2. A minority attended an elite school.
3. most did not complete even their second year.
There are a lot of things that can be read into this, but it does seem to be borne out across most less-regulated industries: founders tend to be smart enough to get into university but perhaps too impatient to stay in for the duration, but leaders of older companies (such as larger and public companies) tend to have at least a four-year degree, although I don't have any hard data about this (particularly because there seems to be a dearth of data for "successful" startup founders outside of the VC world, for any meaning of the word "success".)
Redis is best as an in-memory cache, not a database. Having used it in production for roughly a decade, I don't trust it's on-disk capabilities (AOF/RDB etc) as either solid or reliable (or even performant) in an emergency scenario, especially with DR or DB migration in mind.
It's interesting you say "backup/recovery" as a strong point of relational databases (servers), because backup and recovery on hot databases have always been a challenge.
With many enterprise databases these days, often "incremental" or other seemingly required backup modes are not included in the "community source" versions; perhaps because surely if you want your database to be backed up safely and then come back online safely, you certainly will fall into the "contact us for quote" enterprise customer demographic.
At least, with SQLite, copying even a hot (in-use) db file to a remote server will usually "just work", with the potential loss of a few transactions, but with most other database/servers, you definitely can't just backup the data directory occasionally and call it a day.
Like I mentioned, I don’t have experience working in a start up. My real world experience with backup/recovery of a live relational DB has been with Oracle using ZDLRA - and indeed its license probably costs dearly.
For stuff like MariaDB a quick search also finds options to perform snapshots, backups, restores etc.
And if you need to be super high available, set up a distributed DB like Cassandra - you lose the relational and transaction part, but at least you’re running a product with known failure modes and known ways to prevent/circumvent them.
I guess my bigger point is that besides “don’t roll your own crypto”, I’d also advice not to roll your own DB. There’s a lot of known stuff in the market, all built by people who made and fixed the mistakes you’re going to make a long time ago.
But it was already kinda fringe, even back in the late 90's. People would talk on AIM or AOL chat, but it was rarer to find IRC users among the hoi polloi.
IRC was home to a lot of gaming communities, it was basically Discord before that existed. There used to be live commentary in channels (Twitch replaced this).
I don’t believe any gamers use it today but fringe-tech communities still use it, I think that’s what the parent is referring to.
Yes, their chat is still compatible with IRC, albeit with lots of extensions (which you can opt into using IRCv3 capabilitiy negotiation). More recently they've been adding UI elements on the main site that are web API powered instead of chat though (eg. polls, sending bits).
True but in the early 90s there was nothing else yet. But back then the whole internet was totally fringe of course.
Still, MIRC (as a client) was a common thing for all computer geeks until 2005-2010 or so in my circles. After that people started moving to whatsapp and the like. This includes people in the hacker/maker and game/lanparty community so very techy and niche but there it was still big back then.
