Sun Blogs

It's worth pointing out that this article is a bit long in the tooth. Although perhaps biased responses, it's worth looking at some other thoughts on the subject: http://devcentral.f5.com/weblogs/macvittie/archive/2011/03/21/the-skel… http://dev.robertmao.com/2007/06/30/global-dns-load-balancing-for-free… (towards the end, and yes, the grammar is bad) http://blogs.sun.com/davew/entry/thoughts_on_global_server_load

Being totally new to AWS, why does everyone skip right past using ZFS? http://blogs.sun.com/marchamilton/entry/a_brilliant_argument_for_zfs "Cloud Storage Will Be Limited By Drive Reliability, Bandwidth ... The key feature of ZFS enabling data integrity is the 256-bit checksum that protects your data."

I really enjoyed reading Jonathan's blog at Sun. I'm not sure if he was the first, but he was certainly one of the first large company CEOs to blog and blog candidly (within the constraints of his position). There were a few posts that stuck with me. I remember this one in particular, from 2005, about reducing their customer survey to a single question: would you recommend Sun? http://blogs.sun.com/jonathan/entry/defining_quality I'm glad he's back at it with this new blog.

I suspect they chose SHA256 because it's the best hash, not because it advantages their hardware. But yeah, Niagara 2 is pretty smokin for crypto: http://blogs.sun.com/sprack/entry/ultrasparc_t2_crypto_performance

There are so many (new) languages for the JVM; its becoming a curse of plenty! (Not complaining.) Its a great platform, and invokedynamic will usher in a whole new generation of languages. Lets just hope the Java 7 actually happens given the change of ownership, and I would love to see John Rose's call for programmatic access from Java (http://blogs.sun.com/jrose/entry/jsr_292_support_in_javac) realized. (ASM is great stuff, but I'll admit that its not a great joy to use for me.)

A 2008 Sun article sees Flash as having the following position in the server storage hierarchy: cache, DRAM, Flash, disk, tape: http://blogs.sun.com/jonathan/entry/not_a_flash_in_the Which is what the present article is talking about in this line: The benefits of SSDs as an intermediate caching tier are also limited, and the cost of provisioning such a tier was justified for fewer than 10% of the examined workloads

It depends on how old-school your company is. Play Framework, while it is still using Java, breaks the old-tradition of getter/setter. Your peers might resist instantly when they saw that part. If Play Framework is too extreme for your workplace, I'd suggest to give Spring MVC 3.0 or use the latest Java Servlet 3.0 in JEE6. Java Servlet 3.0 supports annotations (http://blogs.sun.com/swchan/entry/servlet_3_0_annotations) that provides routing capability (might not be as powerful as Rails routing, but still something new and nice to try). For ORM, either JPA 2.0 (recommended) or Hibernate (Hibernate specific APIs, not Hibernate as JPA 2.0 provider). But as with any ORM: you need to know it well (ActiveRecord is not an exception either). Additionally, JPA 2.0 provides Level 2 Cache API interface: you can either write your own but conform with the API (the API is very simple) or use the one provided by the JPA provider. Hibernate (if you use it as the JPA 2.0 provider) supports a variety of L2 caching solutions (SwarmCache, EhCache, OSCache, JBoss Cache, Infinispan, etc). Plug n play baby!

"The decision to go with Microsoft was based on overall value - it was also predicated upon their endorsement of and agreement to help promote MySQL. Stay tuned for more details on what we'll be doing together." From: http://blogs.sun.com/jonathan/entry/winds_of_change_are_blowing This has always been my favourite example of Shwartz' inability to grasp reality. Perhaps I am interpreting this incorrectly but Microsoft "help[ing] promote MySQL?" This deal didn't sound sane in January 2008 and it looks even more ludicrous to me in 2010. By early 2008 the company had lost most of its value despite having a strong brand and a talented pool of engineers. What was going on at Sun? Deals to ship dinky browser bars with Java, JavaFX (which stood no chance in a market dominated by Adobe with Microsoft nipping at their ankles), no roadmap for JAVA 7, inability to see the potential benefits of ZFS on Linux - really, we could be here all evening. You don't "save" a company - You run a company and you should be "saving" it every day with every decision you make.

Copy-on-write is not the same thing as content-addressable. ZFS has deduplication as an optional feature, which is implemented as a content-addressable store of filesystem blocks. In contrast to e.g. git the content-addressable aspect is an implementation detail that is not exposed to users. http://blogs.sun.com/bonwick/entry/zfs_dedup

Thankfully there's a happy fusion of spinning rust and SSD: flashcache [0] and L2ARC/ZIL [1]. Cheaper win. [0]: https://github.com/facebook/flashcache [1]: http://blogs.sun.com/brendan/entry/test

This is relevant: http://blogs.sun.com/bonwick/entry/128_bit_storage_are_you

ZFS supports encryption with deduplication, and Dropbox does dedupe accross customers (and why not? It's not a human digging through things). http://blogs.sun.com/darren/entry/compress_encrypt_checksum_deduplicat… - ZFS link

Without encryption, compression, or deduplication on I don't think there would be very much overheard - more than traditional file systems, definitely, but still relatively very little. Encryption would require significantly more CPU (http://forums.freebsd.org/showthread.php?t=9886) Compression would require moderately more CPU (depends heavily on the compression algorithm) though you can actually speed up disk access with this - which is the real bottleneck (http://blogs.sun.com/observatory/entry/zfs_compression_a_win_win) Deduplication requires lots of memory (I think I saw 1-2GB per TB of data quoted someplace) and will increase CPU load a bit, but for certain workloads can save a lot of writing and diskspace (think Virtual Images or a dataset like Dropbox's). I'm not involved in ZFS development (just an avid user) so take this with a grain of salt.

The lack of a standard for handling namespaces is a fatal flaw in JSON for many applications. Some of the healthcare data exchange formats we deal with require mixing multiple XML namespaces together; it's complex, but it works. Trying to do the same in JSON would be a huge mess. http://blogs.sun.com/bblfish/entry/the_limitations_of_json

Look into what ZFS can do with a spot of flash. ZIL: http://blogs.sun.com/rdm/entry/zil_ssd_and_other_fun L2ARC: http://blogs.sun.com/brendan/entry/test I believe this is the "right way" the article is advocating. You can use it now.

A complete list of JVM options is available online at http://blogs.sun.com/watt/resource/jvm-options-list.html

Not so sure about that. Think about how much energy would be necessary to operate that many devices: http://blogs.sun.com/bonwick/entry/128_bit_storage_are_you

His Nov.2009 blog entry on ZFS data deduplication is a pretty good primer: http://blogs.sun.com/bonwick/en_US/entry/zfs_dedup P.S. also, the 2008 pics with Linus. I suppose there is no further story to go with those ...

Good point, this post talks about their need to rewrite the scrub code to fix a bug and make it faster. http://blogs.sun.com/ahrens/entry/new_scrub_code

This means that Oracle chose Mark Reinhold's plan B (http://blogs.sun.com/mr/entry/rethinking_jdk7). From the point of view of a web application developer, JDK 7 is not very interesting. JDK 7 will contain a few language updates like switch-cases based on strings. No major changes or features that would radically change a web app developer's coding routines. What do you think?

It may not be as flashy in some respects, but things like the time slider are killer http://blogs.sun.com/erwann/entry/zfs_on_the_desktop_zfs It they aren't done already, I suppose integrating these parts to newer versions of Gnome is, at least, in the pipeline for *BSD folks (and, once ported to current Gnome, they would be rather easy to adapt it to BtrFS, which is to be fully supported by canned linux distros by the end of the year). Just now, OpenSolaris is a very cool server OS with a slightly old Gnome desktop on top of it.

Interestingly, this is the same guy that in a discussion of open source economics coined a term for the point when your closed source vendor pushes you too far and you switch to open source and called it the Fuck You, Oracle Point (though there is historical reasons for that name, not just random swearing) http://blogs.sun.com/bmc/entry/the_economics_of_software random quote: "doesn't it strike you as odd that your operating system is essentially free, but your database is still costing you forty grand per CPU?"

You can configure your account to use key based authentication so that you are logged in automatically. Here's an article that I found that explains how. http://blogs.sun.com/jkini/entry/how_to_scp_scp_and

I'm interested to hear Tim's take on it because I remember when he wrote the employee blogging guidelines for Sun Microsystems[1][2]. He was a pretty big force in convincing Sun to talk more candidly in public - up to the CEO himself - so I'm curious to see how his views evolve in the very different environment at Google. [1]: http://blogs.sun.com/lskrocki/entry/sun_s_revised_blogging_policy [2]: http://www.tbray.org/ongoing/When/200x/2004/05/02/Policy

There's a lot of good stuff here--notably the (finally) decent support for literal sets, arrays, and maps in JSON style... but most importantly, closures! I am so sick and tired of creating one-off anonymous classes just to pass a function around. Unfortunately you can only use them with final variables. Ah, well... http://blogs.sun.com/mr/entry/closures http://blogs.sun.com/mr/entry/closures

There are folks within Sun/Oracle doing exactly that. One effort, Project Jigsaw, is aimed at splitting the large rt.jar into smaller modules to reduce the number of classes loaded on startup. Here is an update from last December: http://blogs.sun.com/alanb/entry/is_the_jdk_losing_its Somewhere in there is a comment that says the JVM loads around 300 classes to run an empty main class. I'm curious to see what the merged HotSpot+JRockit JVM will look like in a few years.

Some file systems have this as a built in option. As an example, there's a discussion of the ZFS implementation at http://blogs.sun.com/bonwick/entry/zfs_dedup.

One problem is that in the most lucrative market, enterprise data storage, the woefully out of date benchmarks that are industry standard mainly promote raw throughput on absurdly large working sets, which makes caching of minimal benefit. Besides inertia, there are some established players who have a lot riding on that not changing, also. Details in this kvetch from Sun engineer Bryan Cantrill (Sun's ZFS-based storage servers do pretty much exactly what you propose, but lose out to NetApp in the SPEC benchmarks): http://blogs.sun.com/bmc/entry/eulogy_for_a_benchmark

I can't find the specific link, but here's the marketing pitch: http://www.sun.com/x64/intel/zfs_solution_brief.pdf Also, Googling for "sun hybrid storage" without the quotes will give you more information. Edit: Here's a good explanation of how ZFS incorporates flash in its L2ARC, with benchmarks: http://blogs.sun.com/brendan/entry/test Edit2: I found a higher-level explanation of flash in ZFS. What I didn't realize is that the software is all free to use commercially, though you'd have to be on open solaris: http://blogs.sun.com/jonathan/entry/not_a_flash_in_the

The Black Hole Theory of Design: http://blogs.sun.com/jag/entry/the_black_hole_theory_of Another way to put it is that Lisp might be a local maximum: It may not be the globally optimum language for a particular problem, but once you approach Lisp you find that improving your language requires either becoming more Lisp-like or making a huge leap away from Lisp (C, ML, Haskell, Smalltalk, Erlang, &c.)