Showing posts with label Lustre. Show all posts
Showing posts with label Lustre. Show all posts

Monday, February 11, 2013

Oracle: Solaris 10 Update 11 Released!

Oracle: Solaris 10 Update 11 Released!

Abstract:
Solaris 10 was launched in 2005, with ground-breaking features like: DTrace, SMF (Services), Zones, LDom's, and later ZFS. The latest, and perhaps last, update of Solaris 10 was expected in 2012, to co-inside with an early release of the SPARC T5. In 2013, Oracle released yet another update, suggesting the T5 is close to release. The latest installment of Solaris 10 is referred to as 01/13 release, for January 2013, appears to be the final SVR4 Solaris release, with expected normal Oracle support extending to 2018. Many serious administrators will refer to this release as Solaris 10 Update 11.

(Oracle SPARC & Solaris Road Map, 2013-02-11)

What's New?
Oracle released the "Oracle Solaris 10 1/13 What's New" document, outlining some of the included features. The arrangement of the categories seems odd, in some cases, so a few were merged/re-orded below. Some of the interesting features include:


(Solaris 10 Update 11 Network File System Install Media Option)

(Solaris 10 Update 11 SVR4 Package Dependency Install Support)
  • Administration Enhancements
    OCM (Oracle Configuration Manager) Client Service
    Oracle Zones Pre-Flight Checker
    SVR4 pkgdep (Package Depends) Command
    Intel x86 FMA (Fault Management Architecture) Sandy Bridge EP Enhancements
    AMD MCA (Machine Check Architecture) Support for Family 15h, 0Fh, 10h 
# zfs help                                                              
The following commands are supported:                                   
allow       clone       create      destroy     diff        get         
groupspace  help        hold        holds       inherit     list        
mount       promote     receive     release     rename      rollback    
send        set         share       snapshot    unallow     unmount     
unshare     upgrade     userspace
(Solaris 10 Update 10 zfs help system enhancements)
# zpool help                                                            
The following commands are supported:                                   
add      attach   clear    create   destroy  detach   export   get      
help     history  import   iostat   list     offline  online   remove   
replace  scrub    set      split    status   upgrade                    
# zfs help create                                                       
usage:                                                                  
             create [-p] [-o property=value] ...                        
             create [-ps] [-b blocksize] [-o property=value] ... -V
(Solaris 10 Update 10 zpool help system enhancements)
  • ZFS File System and Storage Enhancements
    Help tiered into sub-commands for: zfs, zpool
    ZFS aclmode enhancements
    ZFS diff enhancements
    ZFS snap alias for snapshot
    Intel x86 SATA (Serial ATA) support for ATA Pass-Through Commands
    AMD x86 XOP and FMA Support
    SPARC T4 CRC32c Acceleration for iSCSI
    Xen XDF (Virtual Block Device Driver) for x86 Oracle VM
# zfs help create                                                       
usage:                                                                  
             create [-p] [-o property=value] ...                        
             create [-ps] [-b blocksize] [-o property=value] ... -V
(Solaris 10 Update 10 zpool help create system enhancements)

Competitive Pressures:
Competition makes the Operating System market healthy! Let's look at the competitive landscape.
(Illumos Logo)

Solaris USB 3.0 is in a better support position than Illumos still missing USB 3.0 today since Solaris 10, Solaris 11, and Illumos all have top-of-the-line read and write flash accelerators for hard disk storage... a USB 3.0 flash cache will provide a nice inexpensive performance boost! Slower Solaris USB 3.0 support from 2013q1 on SPARC will be shunned with Solaris ZFS SMB's considering Apple MacOSX. Apple released USB 3.0 support in 2012q4 with Fusion Drive, making OSX a strong contender. Apple may have been late to Flash when proper licensing could not be agreed between Sun/Oracle and Apple, Apple is still late with deduplication, but now Oracle and Illumos are late with USB 3.0 to combine with ZFS.

(Lustre logo, courtesy hpcwire)

Sun purchased Lustre, for ZFS integration back in 2007. NetMgmt salivated as Lustre for ZFS was on-tap back in 2009, ZFS needed cluster/replication for a long time. Redhat purchased GlusterFS in 2011 and went beta in 2012, for production quality filesystem clustering. IBM released ZFS and Luster on their own hardware & Linux OS. NetMgt noted Lustre on EMC was hitting in 2012, questioned Oracle's sluggishness, and begged for an Illumos rescue. Even Microsoft "got it" when Windows 2012 bundled: dedupe, clustering, iSCSI, SMB, and NFS. It seems Apple, Oracle, and Illumos are the last major vendors - late with native file system clustering... although Apple is not pretending to play on the Server field.

(Superspeed USB 3.0 logo, courtesy usb3-thunderbolt.com)

The lack of File System Clustering in the final update of Solaris 10 is miserable, especially after various Lustre patches made it into ZFS years ago. Perhaps Oracle is waiting for a Solaris 11 update for clustering??? The lack of focus by Illumos on clustering and USB 3.0 makes me wonder whether or not their core supporters (embedded storage and cloud provider) really understand how big of a hole they have. An embedded storage provider, should would want USB 3.0 for external disks and clustering for geographically dispersed storage  their check-list. A cloud provider should would want geographically dispersed clustering, at the least.

(KVM is bundled into Joyent SmartOS, as well as Linux)
Missing native ZFS clustering and hypervisor at Oracle is making Solaris look "long in the tooth". Xen on Oracle Linux with Xen being removed from Solaris is a poor excuse by Oracle. Joyent's SmartOS KVM integrated into Illumos helps the Solaris community move forward, but what is the use of a hypervisor without shared-nothing clustered storage, to migrate those VM's at will? Missing USB 3.0 and native ZFS clustering is putting pressure on Illumos to differentiate itself in the storage market.

Conclusions:
Oracle Solaris 10 is alive and well - GO GET Update 11!!! Some of the most important features include the enhancements to CPU architecture (is SPARC T5 silently supported, since T5 has been in-test since end of 2013?), USB 3.0, iSCSI support for root disk installations, install SVR4 package dependency support, and NFS media support. Many of these features will be welcomed by SMB's (small to medium sized businesses.)

(Bullet Train, courtesy gojapango)
The Solaris Train continues to move at Oracle, producing high quality product, SPARC support, and new drivers (i.e. USB 3.0) - if Solaris 11, Illumos, or SmartOS releases ZFS clustering, the resulting OS will be market leading.
Posted by at No comments:
Labels: , , , , , , , , , , , , , , , , , , ,

Wednesday, July 4, 2012

Clustered File Systems: EMC and Lustre

[EMC VNX HPC Appliance Series, courtesy The Register]

EMC, Lustre, and Massive Storage
EMC announced a VNX appliance with built-in Lustre clustered filesystem storage solution. In an article published by The Register:
The VNX HPC appliance includes a VNX5100 for metadata storage, a VNX7500 for object storage, object and metadata servers, the Terascala LustreStack software and InfiniBand connectivity.
Terascala happens to be a storage start-up located in Massachusetts, just like EMC. EMC, by the way, started shipping Intel based blade servers, pushing Cisco out of their initial partnership. EMC looks

Where is Lustre on Oracle Solaris?
Sun Microsystems purchased Lustre in 2007. Lustre was to be merged onto ZFS. Systems with 256 and fewer nodes could use QFS while 512 nodes would use Lustre. Different updates to OpenSolaris were made to facilitate Lustre integration.

In 2008, Terascala was inquiring with scalability enhancements under ZFS. Oracle purchased Sun and announced support for Lustre 2.x only on Oracle hardware. In 2010, the move of Lustre to Linux distributions such as SUSE seemed inevitable, as Oracle abandoned their support model, and other companies like Clusterstor offered support.

When will Oracle release Lustre with ZFS under Solaris? When will Oracle release native Lustre support on the Oracle storage, as EMC has done?


Open Alternatives
Operating System forks OpenIndiana on source code forks like Illumos, could offer companies such as Terascala, EMC, SI, Xyratex, and such another option: native ZFS merged with Lustre on a base OS which supports all standard protocols: iSCSI, FiberChannel, NFS, CIFS, etc.

This would also solve some of EMC's problems, with needing to find another partner for another cloud project (after dumping Cisco) - they could own the entire cloud, from the hardware (their own blades), to the firmware (VMWare), to the OS (Illumos distribution), to a file system fork (ZFS), cluster fork (Lustre), and all the protocols that go along with Illumos.

When will Illumos release Lustre support?
Posted by at No comments:
Labels: , , , , , , , , , , , ,

Sunday, July 1, 2012

Detecting the Sun in The Solar System

Detecting the the Sun in the Solar System

Abstract:
A question was asked by writer Chris Mellor, after a opinion article in The Register: "What should Oracle do with Sun?" To understand what should be done with Sun, one must understand what Sun's role was in the computing ecosystem, observe the effect of competition, and understand gaps in the market.
Sun's Solar System: Sun Microsystems played in many different overlapping markets sets.

[Image: Sun SPARCStation 1 - courtesy financeportal, The History of the Famous SPARCStation]

Desktop Market
  • Sun commercialized graphical workstations market (may not have created it, but really drove the market) with bundled software applications.
  • Workstations were later killed as Microsoft leveraged it's graphical windows interface to kill off other desktop product offerings, assembling a high-cost portfolio of applications (whose cost were not much different from UNIX workstations) - except UNIX desktops never had an front-office suite.
  • The movement to Ultra-Thin Clients from Sun was an attempt to hold the desktop market, being able to create a new platform, capable of running proprietary Windows applications, as well as Open Standards and Open Sourced applications.
  • Sun's purchase of open-source integrated Office type application and creation of OpenOffice was the attempt to place a foot-hold back in the desktop market (again) - recognizing that Microsoft Office effectively made UNIX workstations irrelevant and Microsoft used the MS Office integrate to drive Microsoft Windows-only based Back Office products to push Solaris out of data centers.
Carrier Server Market
  • Telephone carriers like AT&T used to manufacture their own [Western Electric] 32 bit CPU's for their SVR4 based 3B2 computing systems and use those systems for internal processing.
  • Sun helped to create The Internet, bundling the features that telephone company providers desired, such as SVR4 symmetric multi-processing and standards based interfaces.
  • Desktop Sun Workstations were stacked in racks to make the first clusters, they were re-packaged into rack units. Soon, 32 bit proprietary desktop platforms were re-cased and stacked in racks, and started serving telephone company and internet loads, which were Sun's domain.
  • Pressure was placed on Sun by bundles of open sourced products, as software developers created software on standard-less operating systems (i.e. Linux) on top of proprietary firmware and hardware platforms (which became viable as 16 bit desktop processors gave way to 32 bit processors.)
  • As desktops achieved 64 bit processing, and Windows pushed Sun out of the desktop market, Sun's acquisition SPARC licensee Afara Websystems, and open-sourcing their first processor (for any vendor to share) attempted to push back into web serving loads (where one open-sourced T1 CPU could out-serve as many as 4 or more other proprietary processors.)
  • The use of high-throughput and power conserving carrier-class servers continue to be pumped out of Sun, and later Oracle, in the form of T2, T2+, T3, and finally with the T4 SPARC processors, which were compatible with AT&T SVR4 Solaris.
Enterprise Market
  • The desktop workstation market saw bundled windowing systems, bundled open systems TCP/IP networking, bundled instant messaging (talk), bundled email (SMTP), bundled simple standard text processing (vi, textedit, dtedit, etc.), bundled standard print processing (lp, lpr), added open source simple news group collaboration (NNTP), added open source web server software (Apache), and unlimited user communities - at a reasonable per-unit cost.
  • Companies like Novell and Microsoft released proprietary networking stacks, proprietary operating systems, proprietary desktop environments, per-user licensing fees, costed email clients, with costed email server software, costed print server costs, per machine software charge, etc. Businesses found themselves more likely to build a proprietary software stack of what was available in the market, leveraging UNIX on the back-ends (client-server days.)
  • As desktop providers, like Microsoft, used their operating system as a way to take over the desktop application market, soon it was determined that they could take over the server application market. They released free crippled proprietary desktop products, Server applications, once released upon UNIX (mail servers), would only be available on Windows servers with proprietary protocols. Collaboration tools like News Groups.
  • Java was the "olive branch" offered to the UNIX community, to provide a common software platform, recognizing that proprietary Microsoft Windows effectively made UNIX standards irrelevant and provided an ecosystem for competition with Linux.
Processor Market
  • Sun and other UNIX vendors standardized on real 32 bit platforms with flat memory models, as PC's continued to play with 8 and 16 bit processors.
  • As the proprietary OS desktop wars raged, with proprietary 16 bit processors becoming proprietary 32 bit, Sun and other workstation vendors needed to find a new magic bullet. An Open SPARC consortium was founded, where multiple vendors could cooperate & compete, by implementing their own processors, which complied to a single compatible specification.
  • The inability to bring several open SPARC CPU designs to market placed Sun on the tail-end of general-purpose computing systems for a number of years. This contributed to their demise, and consumption by Oracle
  • Proprietary CPU vendors canceled lines of processors and emulated the radical movement by SPARC, coring-out & threading out, bring their designs to parity in throughput.
  • Sun/Oracle gained time to build a new core, ultimately realized in the SPARC T4, and consolidating more functionality (soon to be realized in the SPARC T5.)


Database Market
The fortunes of Sun Microsystems was increasingly dependent upon Oracle. Oracle had their hand in the decline of Sun Microsystems, by over-milking the cow, and making their platform uncompetitive in the enterprise. For the same performance, Oracle charged a premium under SPARC Solaris, and thus drove the enterprise market off of Solaris.

Sun had engaged into partnerships with Postgres, bundling the product into their Solaris Operating System support, for various ISV's (through which more hardware could be sold.) Sun had also purchased MySQL, to also become a primary ISV vendor support channel (through which more hardware could be sold.) These movements were not necessarily enough, to sway the tide.

Oracle, unfortunately, had their fortunes tied to Sun Microsystem, so Oracle had to buy Sun out, in order to basically survive. The end of Postres support occurred, but Oracle invested more resources in making their database run faster under SPARC Solaris than under any other platform.



Oracle's Education on Sun's History
Interestingly, the new owner, Oracle, can see from the history, that people are more interested in cost than they are in Open or Open Source. If they cared about Open, proprietary desktop applications, proprietary Microsoft Windows, proprietary bios/firmware, and proprietary Intel CPU Architecture would never have eaten Sun's (or other vendors') Open Lunch.

A movement to break into the standard-less Open Source arena was attempted with the creation of OpenSolaris. The new driver was mid-range storage at a terrific price-point. OpenSolaris created new competitors at the low end, in the storage areana, which Oracle did not want to have to wrestle with, so they re-closed Solaris with 11. OpenSolaris did expand Solaris mind share into Open Source arenas.

It is pretty clear what Oracle will do with Sun - Oracle will go proprietary with Sun, to compete with the other proprietary vendors, who ate their Sun's lunch. The market prefers cheap & standardless on proprietary or proprietary on cheap proprietary, or cheap & standardless proprietary on cheap proprietary.

There is value in expensive on proprietary (mixed with standardless open source), if there is enough benefit seen to the consumer, as seen with Apple iPod, iPad, iPhone, etc.



Joyent's Education on Sun's History
Various OpenSolaris distributions have formed around a new Open Source project, where they can continue to share their code contributions upstream to a project greater than their individual entities. OpenSolaris splinters need to realize the move to Cloud is needed.

Kudos to vendors like Joyent. Joyent's effort to consolidate Solaris developers and port KVM (Kernel Virtual Machine) to OpenSolaris was rewarded by Gartner by being recognized in their "Magic Quadrant", when Oracle killed Xen on Solaris.

Illumos's Education on Sun's History
There is limited life expectancy for the OpenSolaris splinters, if all they do is commoditize storage, unless they add something of value. The loss of SVR4 and POSIX features from Illumos as years move ahead is concerning, since they lose the historical value proposition that Solaris offered for carrier providers. Somehow, it appears no one in Illumos is concerned about carriers, but considering that UNIX was created by the carriers - this could be a failure to recall some of their Computer Science 101 history by some of their core developers.

Illumos became lucky with the contributions from Joyent, which keeps them on-par as far as virtualization. Without clustering on ZFS, Oracle and Illumos are both way behind. After Illumos adds file system clustering (they must, to remain relevant as a third-party storage vendor), they could still be at a dead-end, unless they find some other value to contribute to the market (since Linux also has KVM and Illumos is also lacking Xen.)


What Oracle and Illumos Have Not Learned
Various high-end storage vendors have been adding clustered file system, to existing file systems, some have even beat Oracle to bundling a ZFS port with Lustre on a non-Oracle operating system! This places ZFS in Solaris ans Illumos both at great risk or losing relevancy in the Solar System that Sun had carved out for them.

The Oracle Thin-Clients (i.e. SunRay's), being served from the cloud, would be a good start for the Solaris community, since they have something to serve (i.e. Oracle applications, Joyent vitrual desktops, etc.) Illumos still has an opportunity to serve SVR4 POSIX operating system desktops, if they choose to. It is possible that both the Commercial Solaris community with (open SPARC) and Open Source Illumos community (with proprietary Intel) will both decide to cede the ground entirely to Apple, leaving the communities without an ARM presence.

Concluding Thoughts:
The market is still pretty much defined by the work that AT&T and Sun had done, historically. If Oracle and Illumos abandon their positions, the known universe will continue to degrade from standards-based systems into proprietary commodity based solutions - which may be the new universe what Oracle desires to play in. Where this leaves the rest of the market, that is a good question, but The Sun has not collapsed into a black-hole yet - it is still quite observable, no matter how much people continue to try to ignore it in the sky.
Posted by at No comments:
Labels: , , , , , , , , , ,

Wednesday, June 27, 2012

The Processor Market: POWER #1 HPC

The Processor Market: POWER #1 HPC

Abstract:
During June 2012, some very interesting updates happened - some Open Source pieces from Sun and Oracle were combined with the POWER processors to build a new Super Computer. An odd result: IBM's POWER required a lot more sockets to outrun Fujitsu SPARC64... but did so with better power efficiency and using arch-rival Sun Microsystem's (now Oracle's) open source technology.

 
[wiring 123% more sockets for 55% greater performance, courtesy The Register]
IBM Denies Fujitsu's SPARC64 Year Long #1 HPC Rank!
With a long list of losses, IBM's POWER architecture finally has a win: proprietary IBM POWER architecture now has a #1 HPC Performance spot with Lustre under ZFS - denying Fujitsu their nearly 1 year long spot as the fastest computer in the world, with Fujitsu's fork of Lustre clustered filesystem!

[zfs write performance under linux with lustre, courtesy Lawrence Livermore Laboratory]
Whamcloud, Lustre, and Sequoia Supercomputer

The Lustre clustered/distributed filesystem, formerly owned by Sun Microsystems, now owned by Oracle. It has long been promised to be merged into ZFS. Whamcloud is a commercial enterprise which develops a fork of the Lustre file system. They announced the release of Chroma Enterprise, to bring enterprise management to Lustre.

Whamcloud is using a non-kernel emulated ZFS fork from OpenSolaris. The ZFS implementation still shows linear scalability (in comparison to the native Linux filesystem), as the load increases.

The Sequoia Supercomputer, run by the United States Department of Energy, has an interesting feature - the use of a merged Sun's  ZFS and Sun's Lustre filesystem. Here is a short 30 minute video talking to the PDF from the Lustre User Group (LUG) 2012.

IBM's Tortoise vs Fujitsu's Hare
IBM needed 123% more proprietary POWER CPU sockets to outrun Fujitsu's open SPARCv9 SPARC64 architecture by a mere 55%. The IBM POWER solution proved itself to be about 23% more power efficient, which is truly an achievement, considering how many more sockets were required. The tortoise POWER processor takes less energy than the hare SPARC64 processor.

Fujitsu SPARC64 Loses The Battle of the Alamo
This is somewhat a Pyrrhic victory, kind of like winning the Battle of the Alamo. Could any 1 year old platform hold it's performance position, when the new opposition has a 123% numeric advantage?

This victory was a solid win for IBM, from a supercomputer to supercomputer perspective, but there is an odd conclusion that some people may notice: each SPARC64 old socket appears to demonstrate a minimum of 123% faster than each new POWER socket.

Considering that each SPARC64 socket was an 8 core processor socket, in comparison to the 18 core POWER processor socket (of which 16 cores is usable) - each SPARC64 core is roughly 243% faster than each POWER core!

Fujitsu's SPARC64 Other Battle FrontsThe battles have been continuous since 2011:
SPARC continue to be on the map, in new locations, as well as eating IBM POWER's lunch in smaller installations - for very good reason. The new 16 core SPARC64 chips offer double the performance, in the same socket, making POWER look pale, in comparison.


Better Options for Super Computers
IBM's main processor is POWER with it's main OS being AIX. AIX is lacking a modern file system. IBM had a second operating system option, Linux, but it was lacking a modern file system. IBM briefly toyed with the idea of purchasing Sun Microsystems, before Oracle made the final purchase. AIX and Linux choices on POWER were lacking.

Why was the choice made to emulate ZFS? The licensing in Linux is so restrictive that ZFS could not be combined with the Linux kernel, so it had to be emulated in userland. Why did IBM use Lustre instead of IBM's own GPFS clustered file system? Cost may be a factor and Lustre is basically the defacto standard in High Performance Computing.

Lustre was going to be merged into ZFS by Sun Microsystems, after it's acquisition in 2007. The use of Lustre support directly from Oracle, without hardware, came to an end shortly after the purchase of Sun Microsystems by Oracle. Oracle limited the support of Lustre to Oracle hardware in 2010.

Code changes to OpenSolaris were delivered for Lustre friendliness - the movement to complete Lustre with ZFS under Illumos in kernel space could have offered better performance over user space ZFS, fewer system calls would be required at the emulation layer. Illumos could have delivered native performance on the IBM POWER Sequoia or the Fujitsu SPARC64 K Supercomputer.

Fujitsu, being the SPARC64 creator, was more than capable of delivering their drivers into the Illumos market, had Illumos been interested in SPARC. Clearly, pushing IBM to adopt forks of Oracle's Solaris ZFS and Oracle's Lustre was still pretty aggressive, perhaps pushing them all the way to adopt Illumos, a fork of Solaris, was a bridge too far (especially, after a failed Solaris acquisition.)
Conclusions
With some in the Illumos community seemingly less interested in POSIX subsystems, pulling out SVR4 features, disinterested in non-Intel distributions - some are asking the question the value of Illumos without the differentiators of ZFS and DTrace with an OS like Linux.

With POWER sitting as #1, SPARC64 as #2, and ARM growing with increasing market prevalence - the window for Illumos relevance may be closing if they don't start actively supporting some non-x64 architectures, as their differentiating features get ported to competing OS's.

IBM's POWER has long tried to demonstrate their superiority in per-socket or per-core performance. The POWER platform uses 18 core's per socket while Fujitsu uses 8 cores per socket - so each POWER core is vastly slower than a Fujitsu SPARC64 core.

IBM long tried to demonstrate their superiority of technologies to companies like Sun and Oracle, yet at the core of their super computer was ZFS and Lustre - in order to compete in this arena, former Sun Microsystem (now Oracle) technology was used, to scale their solution.

A non-IBM operating system, running a fork of Oracle Solaris ZFS, and running a fork of Oracle Lustre is not the way some might want to advertize an IBM POWER architecture (which normally runs IBM AIX operating system with IBM GPFS file system.)
Posted by at No comments:
Labels: , , , , , , , ,

Wednesday, October 7, 2009

ZFS: The Next Word

ZFS: The Next Word

Abstract

ZFS is the latest in disk and hybrid storage pool technology from Sun Microsystems. Unlike competing 32 bit file systems, ZFS is a 128-bit file system, allowing for near limitless storage boundaries. ZFS is not a stagnant architecture, but a dynamic one, where changes are happening often to the open source code base.

What's Next in ZFS?

Jeff Bonwick and Bill Moore did a presentation at The Kernel Conference Australia 2009 regarding what was happening next in ZFS. A lot of the features were driven by the Fishworks team as well as Lustre clustering file system.

What are the new enhancements in functionality?
  • Enhanced Performance
    Enhancements all over the system
  • Quotas on a per-user basis
    Always had quotas on a per-filesystem basis, originally thought each user would get a filesystem, this does not scale well for thousands of users with many existing management tools
    Works with industry standard POSIX based UID's & Names
    Works with Microsoft SMB SID's & Names
  • Pool Recovery
    Disk drives often "out-right lie" to operating system when they re-order the writing of the blocks.
    Disk drives often "out-right lie" to operating systems when they receive a "write barrier", indicating that the write was completed, when the write was not completed.
    If there is a power outage in the middle of the write, even after a "write barrier" was done, the drive will often silently drop the "write commit", making the OS thinking that the writes were safe, when they were not - resulting in a pool corruption.
    Simplification in this area - during a scrub, go back to an earlier uber-block, and correct pool... and never over-write a recently changed transaction group, in the case of a new transaction.
  • Triple Parity RAID-Z
    Double parity RAID-Z has been around from the beginning (i.e. lose 2 out of 7 drives)
    Triple parity RAID-Z allows for disks with bigger, higher, faster high-BER drive usage
    Quadruple Parity is on the way (i.e. lose 3 out of 10 drives)
  • De-duplication
    This is very nice capacity enhancement with application, desktop, and server virtualization
  • Encryption
  • Shadow Migration (aka Brain Slug?)
    Pull out that old file server and replace it with a ZFS [NFS] server without any downtime.
  • BP Rewrite & Device Removal
  • Dynamic LUN Expansion
    Before, if a larger drive was inserted, the default behavior was to resize the LUN
    During a hot-plug, tell the system admin that the LUN has been resized
    Property added to make LUN expansion automatic or manual
  • Snapshot Hold property
    Enter an arbitrary string for a tag, issue the snapshot, issue a delete, when an "unhold" is done, the destroy is done.
    Makes ZFS look sort of like a relational database with transactions.
  • Multi-Home Protection
    If a pool is shared between two hosts, works great as long as clustering software is flawless.
    The Lustre team prototyped a heart-beat protocol on the disk to allow for multi-home-protection inherent in ZFS
  • Offline and Remove a separate ZFS Log Device
  • Extend Underlying SCSI Framework for Additional SCSI Commands
    SCSI "Trim" command, to allow ZFS to direct less wear leveling on unused flash areas, to increase life and performance of flash
  • De-Duplicate in a ZFS Send-Receive Stream
    This is in the works, to make backups & Restores more efficient
Performance Enhancements include:
  • Hybrid Storage Pools
    Makes everything go (alot) faster with a little cache (lower cost) and slower drives (lower cost.)
    - Expensive (fast, reliable) Mirrored SSD Enterprise Write Cache for ZFS Intent Logging
    - Inexpensive consumer grade SSD cache for block level Read Cache in a ZFS Level 2 ARC
    - Inexpensive consumer grade drives with massive disk storage potential with a 5x lower energy consumption
  • New Block Allocator
    This was a extremely simple 80 line code segment that works well under empty pools, that was finally re-engineered for performance when the pool gets full. ZFS will now use both algorithms.
  • Raw Scrub
    Increase performance by running through the pool and metadata to ensure checksums are validated without uncompressing data in the block.
  • Parallel Device Open
  • Zero-Copy I/O
    From the folks in Lustre cluster storage group requested and implemented the feature.
  • Scrub Prefetch
    A scrub will now prefetch blocks to increase utilization of the disk and decrease scrub time
  • Native iSCSI
    This is part of the COMSTAR enhancements. Yes, this is there today, under OpenSolaris, and offers tremendous performance improvements and simplified management
  • Sync Mode
    NFS benchmarking in Solaris is shown to be slower than Linux, because Linux does not guarantee a write to NFS actually makes it to disk (which violates the NFS protocol specification.) This feature allows Solaris to use a "Linux" mode, where writes are not guaranteed, to increase performance, at the expense of .
  • Just-In-Time Decompression
    Prefetch hides latency of I/O, but burns CPU. This allows prefetch to get the data without decompressing the data, until needed, to save CPU time, and also conserve kernel memory.
  • Disk drives with higher capacity and less reliability
    Formatting options to reduce error-recovery on a sector-by-sector basis
    30-40% improved capacity & performance
    Increased ZFS error recovery counts
  • Mind-the-Gap Reading & Writing Consolidation
    Consolidate Read Gaps in the case of reads, to ingle aggregate read can be used, reading data between adjacent sectors, and throw away intermediate data, since fewer I/O's allow for streaming data from drives more efficiently
    Consolidate Write Gaps in the case of a write, so single aggrigate write can be used, even if adjacent regions have a blank sector gap between them, streaming data to drives more efficiently
  • ZFS Send and Receive
    Performance has been improved using the same Scrub Prefetch code
Conclusion

The ZFS implementation in Solaris 10-2009 release actually has some of the ZFS features detailed in the most recent conferences.
Posted by at No comments:
Labels: , , , , , , ,
Subscribe to: Posts (Atom)