Its not only about performance, Redis data structures offer an even more advanced caching and data processing. I even use Redis as a cache for ClickHouse.
> I would rather pay a competent cloud provider than being responsible for reliability issues.
Why do so many developers and sysadmins think they're not competent for hosting services. It is a lot easier than you think, and its also fun to solve technical issues you may have.
The point was about redundancy / geo spread / HA. It’s significantly more difficult to operate two physical sites than one. You can only be in one place at a time.
If you want true reliability, you need redundant physical locations, power, networking. That’s extremely easy to achieve on cloud providers.
You can just rent the rack space in datacenter and have that covered. It's still much cheaper than running that in cloud.
It doesn't make sense if you only have few servers, but if you are renting equivalent of multiple racks of servers from cloud and run them for most of the day, on-prem is staggeringly cheaper.
We have few racks and we do "move to cloud" calculation every few years and without fail they come up at least 3x the cost.
And before the "but you need to do more work" whining I hear from people that never did that - it's not much more than navigating forest of cloud APIs and dealing with random blackbox issues in cloud that you can't really debug, just go around it.
On cloud it's out of your control when an AZ goes down. When it's your server you can do things to increase reliability. Most colos have redundant power feeds and internet. On prem that's a bit harder, but you can buy a UPS.
If your head office is hit by a meteor your business is over. Don't need to prepare for that.
> Why do so many developers and sysadmins think they're not competent for hosting services.
Because those services solve the problem for them. It is the same thing with GitHub.
However, as predicted half a decade ago with GitHub becoming unreliable [0] and as price increases begin to happen, you can see that self-hosting begins to make more sense and you have complete control of the infrastructure and it has never been more easier to self host and bring control over costs.
> its also fun to solve technical issues you may have.
What you have just seen with coding agents is going to have the same effect on "developers" that will have a decline in skills the moment they become over-reliant on coding agents and won't be able to write a single line of code at all to fix a problem they don't fully understand.
> Why do so many developers and sysadmins think they're not competent for hosting services. It is a lot easier than you think, and its also fun to solve technical issues you may have.
It is a different skillset. SRE is also an under-valued/paid (unless one is in FAANGO).
Maybe you find it fun. I don’t, I prefer building software not running and setting up servers.
It’s also nontrivial once you go past some level of complexity and volume. I have made my career at building software and part of that requires understanding the limitations and specifics of the underlying hardware but at the end of the day I simply want to provision and run a container, I don’t want to think about the security and networking setup it’s not worth my time.
Because when I’m running a busy site and I can’t figure out what went wrong, I freak out. I don’t know whether the problem will take 2 hours or 2 days to diagnose.
At a previous job, the company had its critical IT infrastructure on their own data center. It was not in the IT industry, but the company was large and rich enough to justify two small data centers. It notably had batteries, diesel generators, 24/7 teams, and some advanced security (for valid reasons).
I agree that solving technical issues is very fun, and hosting services is usually easy, but having resilient infrastructure is costly and I simply don't like to be woken up at night to fix stuff while the company is bleeding money and customers.
The biggest advantage of ZFS from a operational experience, is that when you have problems, ZFS tells you why. Checksum errors? Something wrong with the hard drive or SATA/SAS cables. Is the disk slow, zfs events will tell you that it spent more than 5 seconds to read sector x from disk /dev/sdf. The zfs cli commands are super-intuitive, and makes fully sense. Compared to ie. virsh, which is just weird for managing vm's.
It definitely worth the hassle. But if everything works fine for you now, don't bother. ZFS is not going away and you can learn it later.
It’s been 17 years since I got my first Linux job in 2008. Where I live, that’s rare, 99% of the industry here is a 'Microsoft Shop,' and the biggest player in town is practically married to them.
I started out at a small Linux company working with Plone CMS. The pay wasn’t great, but it was the perfect place to learn Linux and Python. Since then, I’ve used Linux every single day, become a Java developer, and started a few businesses. Using Linux of course.
But lately, things are changing. Companies are realizing that when it comes to Data Engineering and Science, C# just can't compete with Python's ecosystem. Now that they need to pivot, they're looking for help, and there are very few people in this area with the experience to answer that call.
Do not buy Kioxia. If your drives dies after a few years, but can come back to life with an firmware update. They will not give it to you unless you have a support contract.
Solidigm have all their firmware available for everyone their website.
As a startup, what is it in for me to switch from Java, Spring Boot, Hibernate, Beam, Flink, Pulsar, Vault, KeyCloak ecosystem to C#.Net? Is the documentation better? Do I get better performance? Is the community larger and more stable?
As others have mentioned Vault, Keycloak, Flink are language agnostic. Regarding the switch from Java to .NET, I would rather recommend switching to Kotlin instead of .NET for a developer experience similar to C#, while still keeping your existing expertise in Java and its ecosystem.
And this comes from someone in a .NET shop currently, but have worked with Java before.
IMHO both languages and surrounding ecosystems are good. Both have their pros and cons and quirks.
Most of that ecosystem is language agnostic, or offer much more ergonomically sane APIs in dotnet. This is especially true for anything coming out of Google (e.g. Dataflow which runs on top of Apache Beam).
C# itself has way better DX (object initializers alone are worth the switch), and most language features don't feel bolted on like with Java (anything from functional programming to extension methods to whatever).
And at least 6 years ago .net with default settings required significantly less resources (RAM, CPU) and yad significantly faster startup than comparable Java code.
C# is also significantly more consistent. You might not use LINQ, but since everything is IEnumerable, you will use the same set of methods on everything. None of the Lis.of...Collectors.collect idiocy from Java.
I also found Asp.net to have significantly less undebuggable magic than Spring.
I sometimes miss Spring magic when working with ASP.NET, and I worked 12+ years with C# and only a year with Spring. Not saying one is better than the other, it's always a choice, less magic = more boilerplate and less boilerplate = more magic.
> and most language features don't feel bolted on like with Java (anything from functional programming to extension methods to whatever)
Java doesn't have extension methods and while both are decent languages, C# is the one that likes implementing every conceivable language feature immediately, while Java takes a while to design a bigger feature that will replace several smaller ones' use cases.
> while Java takes a while to design a bigger feature that will replace several smaller ones' use cases.
Java is quite busy also implementing features that are small in other languages like text blocks.
And I wish Java would design bigger features that would replace several "smaller ones", but that is almost never the case. It's almost always just a new big feature bolted on to the language that is almost there, but not quite
There are many reasons for that, none of them simple, and it doesn't help that there's also the attitude of "those lesser languages cannot compare to the greatness that is Java" [1]
If we're talking about Java, somehow you're still required to do builder patterns and manually create `.of` constructors for everything. Where C# has had object initialisers and pervasive IEnumerable with a very simple interface that nearly everything uses. And that is only scratching the surface of DX.
Almost every feature bolted onto Java seems to take several times more code, and doesn't really work with the rest of the language.
Don't get me wrong, I quite like working with modern Java. But I had the chance to work at a company where micro services where developed in both Java and C#, and the difference is still light and day.
> Java is quite busy also implementing features that are small in other languages like text blocks.
It's a small feature which is immediately understandable by anyone coming straight from Java 1.2, it doesn't materially increase the complexity of the language and is arguably one of the best implementation among different languages. So not really sure if it's a good counterpoint.
Meanwhile records arrived several versions ago with sealed interfaces/classes on the horizon already, so now they together form a complete ADT feature. Pattern matching builds on top. And sure, these are no novel features, MLs had this decades ago, but the implementation is very nice, with minimal additional developer complexity and some small DX improvement (records replace the majority of Bean usages). In the future, withers may come that would help with both object initialization AND record "mutation". Where Java spends more development budget is on the runtime side, e.g. virtual threads can replace async code in many cases, so the language doesn't have to get all the complexity of an async feature.
Meanwhile C# has many many "one-off" features and they really can have unexpected interactions and make the language quite a bit harder to understand. Some of them are absolutely wonderful, but I am on a "the-smaller-the-merrier" language team.
> It's a small feature which is immediately understandable by anyone coming straight from Java 1.2
You literally started with "Java takes a while to design a bigger feature that will replace several smaller ones' use cases".
But look, this small feature is not a one-off, and is actually useful because <arbitrary arguments>
> Meanwhile records arrived several versions ago
> Meanwhile C# has many many "one-off" features
Such one-off features like records and pattern-matching that C# also has? Or multiline strings that C# also has?
Or any other useful features directly impacting developer experience that you just dismiss out of hand because for some arbitrary reason you deem them "one-off" or small. Until they finally make their way into Java of course.
Vault, Keycloak, Flink are language agnostic or there exist bindings for most popular languages.
Documentation is vastly better compared to Java ones, it's like day and night, LINQ is vastly superior to anything that Java offered - but i haven't used java in a very long time. And every time i had to write java it felt like i went backwards in time by 5-10 years.
If i remember right Java's webserver beats ASP.NET in performance benchmarks but .net's one performance is good enough that it does not matter until you hit really big usercount - and at that point you usually have to rethink your architecture anyways.
But frankly .net is still mostly Microsoft Java but with better developer ergonomics in my opinion. It did shed a lot of overengineered OOP legacy from .net framework days though and we're seeing major performance improvements with every version.
What was the last Java version you used? There has been a huge momentum in adding new features lately, granted, it is slower than in C# (Java's top priority is backwards compatibility, so it does not have the luxury of shedding old stuff or changing them once they are in), but in the last couple of years it has improved tremendously. The JVM (especially in the garbage collection front) but also the language - half of an ML-style language is there (for example, ADTs and pattern matching), the other half is coming soon!
currently according to techempower benchmarks ASP.net is 55th overall in minimal variant, while being 83 in normal one in Fortunes benchmark which is basically a normal usecase.
Look at plaintext results if you want to compare just servers. Fortunes benchmark has too many variables (including db queries etc) to say anything about server performance. As for fortunes benchmark, for more realistick Java performance numbers I would look at Spring. Also there are many shady things on those Java bencmarks.
I think saying that Spring is the representative of Java metrics is somewhat equivalent to saying that full aspnet mvc is the representative of dotnet metrics.
On the dotnet side, both Oxpecker and Giraffe (Giraffe being written by the author of that post) perform very well with simple code and from what I see, no "tricks". It's all standard "how the docs say to write it" code (muuuuch different than those platform benchmarks that were rightfully scrutinized in the referenced blog post).
On the jvm side, I started looking for a reference near the top without any targeted non-default optimizations (which is really what I personally look for in these). The inverno implementation has a few things that I'd call non-standard (any time I see a byte buffer I imagine that's not how most people are utilizing the framework), but otherwise looks normal. I recall an earlier quarkus implementation that I read through a couple years ago (same repo) that wasn't as optimized with small things like that and performed very well, but it seems they've since added some of those types of optimizations as well.
All to say: If you venture outside the standard of either platform (full fatty aspnet/ef or spring/hibernate) you can make the tradeoff of framework convenience for performance. However when it comes to the cost/benefit ratio, you're either going to be joining a company using the former, or writing your own thing using the latter (most likely).
I do not have any benchmarks other than this[1] to refer to, but I work with Quarkus[2] and Java 25 LTS (just recently released) services deployed on AWS EKS and we are very happy with the performance (for mobile game backends)
Quarkus does a lot of bytecode generation magic at build time, which will give it an "unfair" edge in some scenarious, like this simple serialization/deserialization case in this particular benchmark.
Look at plaintext benchmark, if you want to compare just servers. Also look at Spring score in fortunes, which is the more common Java stack and I think a more suitable comparison.
> Documentation is vastly better compared to Java ones, it's like day and night
This is absolutely not my experience, especially when it comes to the ecosystem and third-party libraries. Like Java is pretty much the best in this category.
reply