A client asked me to explain "Nearline storage" to them. This was easy, I thought, as I started my IBM career on DFHSM, now known as DFSMShsm for z/OS, which was created in 1977 to support the IBM 3850 Mass Storage System (MSS), a virtual storage system that blended disk drives and tape cartridges with robotic automation. Here is a quick recap:
- Online storage is immediately available for I/O. This includes DRAM memory, solid-state drives (SSD), and always-on spinning disk, regardless of rotational speed.
- Nearline storage is not immediately available, but can be made online quickly without human intervention. This includes optical jukeboxes, automated tape libraries, as well as spin-down massive array of idle disk (MAID) technologies.
- Offline storage is not immediately available, and requires some human intervention to bring online. This can include USB memory sticks, CD/DVD optical media, shelf-resident tape cartridges, or other removable media.
These terms and their definitions have been used for decades, and are consistent with or at least similar to definitions I found on [Wikipedia], [Webopedia], [WiseGEEK], and [SearchStorage].
Sadly, it appears a few storage manufacturers and vendors have been misusing the term "Nearline" to refer to "slower online" spinning disk drives. I find this [June 2005 technology paper from Seagate], and this [2002 NetApp Press Release], the latter of which included this contradiction for their "NearStore" disk array. Here is the excerpt:
"Providing online access to reference information—NetApp nearline storage solutions quickly retrieve and replicate reference and archive information maintained on cost-effective storage—medical images, financial models, energy exploration charts and graphs, and other data-intensive records can be stored economically and accessed in multiple locations more quickly than ever"
Which is it, "online access" or "nearline storage"?
If a client asked why slower drives consume less energy or generate less heat, I could explain that, but if they ask why slower drives must have SATA connections, that is a different discussion. The speed of a drive and its connection technology are for the most part independent. A 10K RPM drive can be made with FC, SAS or SATA connection.
I am opposed to using "Nearlne" just to distinguish between four-digit speeds (such as 5400 or 7200 RPM) versus "online" for five-digit speeds (10,000 and 15,000 RPM). The difference in performance between 10K RPM and 7200 RPM spinning disks is miniscule compared to the differences between solid-state drives and any spinning disk, or the difference between spinning disk and tape.
I am also opposed to using the term "Nearline" for online storage systems just because they are targeted for the typical use cases like backup, archive or other reference information that were previously directed to nearline devices like automated tape libraries.
Can we all just agree to refer to drives as "fast" or "slow", or give them RPM rotational speed designations, rather than try to incorrectly imply that FC and SAS drives are always fast, and SATA drives are always slow? Certainly we don't need new terms like "NL-SAS" just to represent a slower SAS connected drive.
technorati tags: IBM, online, nearline, offline, FC, SATA, SAS, NL-SAS, MAID, SSD, DVD, optical, NetApp, Seagate,
Has EMC stooped so low that they have to resort to Hitachi math for their latest performance claims?
Readers might remember that just a few months ago, I had a blog post [Is this what HDS tells our mainframe clients?] pointing out the outlandish comparison Hitachi was using in their presentations. Their response was to cover it up, forcing me to follow up with my post [The Cover-up is worse than the original crime]. To their credit, they eventually removed the false and misleading information from their materials.
Now an avid reader of my blog has brought this to my attention. Apparently,
EMC has been showing customers a presentation
[Accelerating Storage Transformation with VMAX and VPLEX] with false and misleading comparison claims between IBM DS8000, HDS VSP and EMC VMAX 40K disk system performance.
(FTC Disclosure: This would be a good time to remind my readers that I work for IBM and own IBM stock. I do not endorse any of the EMC or HDS products mentioned in this post, and have no financial affiliation or investments directly with either EMC nor HDS. I am basing my information solely on the presentation posted on the internet and other sources publicly available, and not on any misrepresentations from EMC speakers at the various conferences where these charts might have been shown.)
The problem with misinformation is that it is not always obvious. The EMC presentation is quite pretty and professional-looking. It is the typical slick, attention-getting, low-content, over-simplified marketing puffery you have come to expect from EMC. There are two slides in particular that I have issue with.
This first graphic implies that IBM and HDS are nearly tied in performance, but that EMC VMAX 40K has nearly triple that bandwidth. Overall the slide has very little detail. That makes it difficult to determine what exactly is being claimed and whether a fair comparison is being made.
- The title claims that VMAX 40K is "#1 in High Bandwidth Apps". Only three disk systems are shown so the claim appears to be relative to only the three systems. The wording "High Bandwidth Apps" is confusing considering the cited numbers are for disk systems and no application is identified. By comparison, IBM SONAS can drive up to 105 GB/sec sequential bandwidth, nearly double what EMC claims for its VMAX 40K, so EMC is certainly not even close to #1.
- Is the workload random or sequential? That is not easy to determine. The use of "GB/s" along with the large block size of 128KB implies the I/O workload is sequential, which is great for some workloads like high performance computing, technical computing and video broadcasts. Random workloads, on the other hand, are usually measured in I/Os per second (IOPS) with a block size ranging 4KB to 64KB. (I am assuming the 128K blocks refers to 128KB block size, and not reading the same block of cache 128,000 times.)
- The slide states "Maximum Sustainable RRH Bandwidth 128K Blocks". The acronym "RRH" is not defined; but I suspect this refers to "random read hits". For random workloads, 100 percent random read hits from cache represents one corner of the infamous "four corners" test. Real-world workloads have a mix of reads and writes, and a mix of cache hits and cache misses. It is also unclear whether the hits are from standard data cache or from internal buffers in adapters (perhaps accessing the same blocks repeatedly) or something else. So is this really for a random workload, or a sequential workload?
(The term "Hitachi Math" was coined by an EMC blogger precisely to slam Hitachi Data Systems for their blatant use of four-corners results, claiming that spouting ridiculously large, but equally unrealistic, 100 percent random read hit results don't provide any useful information. I agree. There are much better industry-standard benchmarks available, such as SPC-1 for random workloads, SPC-2 for sequential workloads, and even benchmarks for specific applications, that represent real-world IT environments. To shame HDS for their use of four-corners results, only for EMC themselves to use similar figures in their own presentation is truly hypocritical of them!)
- The IBM system is identified as "DS8000". DS8000 is a generic family name that applies to multiple generations of systems first introduced in 2004. The specific model is not identified, but that is critical information. Is this a first generation DS8100, or the latest DS8800, or something in between?
The slide says "Full System Configs", but that is not defined and configuration details are not identified. Configuration details, also critical information in assessing system performance capabilities, are not specified. If the EMC box costs seven times more than IBM or HDS, would you really buy it to get 3x more performance? Is the EMC packed with the maximum amount of SSD? Were there any SSD in the IBM or HDS boxes to match?
The source of the claimed IBM DS8000 performance numbers is not identified. Did they run their own tests? While I cannot tell, the VMAX may have been configured with 64 Fibre Channel 8Gbps host connections. In that case each channel is theoretically capable of supporting about 800 MB/s at 100% channel utilization. Multiplying 64 x 800MB/s = 51.2GB/s, so did EMC just do the performance comparison on the back of a napkin, assuming there are no other bottlenecks in the system? Even then, I would not round up 51.2 to 52!
- Response times were not identified. For random I/Os, response time is a very important metric. It is possible that the Symmetrix was operating with some resources at 100% utilization to get the highest GB/s result, but that would likely make I/O response times unacceptable for real-world random I/O workloads.
IBM and HDS have both published Storage Performance Council [SPC] industry-standard performance benchmarks. EMC has not published any SPC benchmarks for VMAX systems. If EMC is interested in providing customers with audited, detailed performance information along with detailed configuration information, all based on benchmarks designed to represent real-world workloads, EMC can always publish SPC benchmark results as IBM and other vendors have done. In past blog fights, EMC resorts to the excuse that SPC isn't perfect, but can they really argue that vague and unrealistic claims cited in its presentation are better?
The second graphic is so absurd, you would think it came directly from Larry Ellison at an Oracle OpenWorld keynote session. EMC is comparing a configuration with VMAX 40K plus an EMC VFCache host-side flash memory cache card to a configuration with an IBM and HDS disk system without host-side flash memory cache also configured. The comparison is clearly apples-to-oranges. Other disk system configuration details are also omitted.
FAST VP is EMC's name for its sub-volume drive tiering feature, comparable to IBM Easy Tier and Hitachi's Dynamic Tiering. The graph implies that IBM and HDS can only achieve a modest increment improvement from their sub-volume tiering. I beg to differ. I have seen various cases where a small amount of SSD on IBM DS8000 series can drastically improve performance 200 to 400 percent.
The "DBClassify" shown on the graph is a tool run as part of an EMC professional services offering called Database Performance Tiering Assessment, makes recommendations for storing various database objects on different drive tiers based on object usage and importance. Do you really need to pay for professional services? With IBM Easy Tier, you just turn it on, and it works. No analysis required, no tools, no professional services, and no additional charge!
- VFCache is an optional product from EMC that currently has no integration whatsoever with VMAX. A fair comparison would have included a host-side flash memory cache (from any vendor) when the IBM or HDS storage system was configured. Or leave it out altogether and just focus on the sub-volume tiering comparison.
Keep in mind that EMC's VFCache supports only selected x86-based hosts. IBM has published a [Statement of Direction] indicating that it will also offer this for Power systems running AIX and Linux host-side flash memory cache integrated with DS8000 Easy Tier.
I feel EMC's claims about IBM DS8000 performance are vague and misleading. EMC appears to lack the kind of technical marketing integrity that IBM strives to attain.
Since EMC is not able or willing to publish fair and meaningful performance comparisons, it is up to me to set the record straight and point out EMC's failings in this matter.
Reminder: It's not to late to register for my Webcast "Solving the Storage Capacity Crisis" on Tuesday, September 25. See my blog post [Upcoming events in September] to register!
technorati tags: IBM, DS8000, DS8800, HDS, VSP, EMC, VMAX, Symmetrix, VFCache, Easy Tier, FAST VP
The technology industry is full of trade-offs. Take for example solar cells that convert sunlight to electricity. Every hour, more energy hits the Earth in the form of sunlight than the entire planet consumes in an entire year. The general trade-off is between energy conversion efficiency versus abundance of materials:
- Get 9-11 percent efficiency using rare materials like indium (In), gallium (Ga) or cadmium (Cd).
- Get only 6.7 percent efficiency using abundant materials like copper (Cu), tin (Sn), zinc (Zn), sulfur (S), and selenium (Se)
IBM has eliminated this trade-off with a record-setting breakthrough last week, demonstrating 9.6 percent efficiency [thin film solar cells using earth-abundant materials].
A second trade-off is exemplified by EMC's recent GeoProtect announcement. This appears similar to the geographic dispersal method introduced by a company called [CleverSafe]. The trade-off is between the amount of space to store one or more copies of data and the protection of data in the event of disaster. Here's an excerpt from fellow blogger Chuck Hollis (EMC) titled ["Cloud Storage Evolves"]:
"Imagine a average-sized Atmos network of 9 nodes, all in different time zones around the world. And imagine that we were using, say, a 6+3 protection scheme.
The implication is clear: any 3 nodes could be completely lost: failed, destroyed, seized by the government, etc.
-- and the information could be completely recovered from the surviving nodes."
For organizations worried about their information falling into the wrong hands (whether criminal or government sponsored!), any subset of the nodes would yield nothing of value -- not only would the information be presumably encrypted, but only a few slices of a far bigger picture would be lost.
Seized by the government? falling into the wrong hands? Is EMC positioning ATMOS as "Storage for Terrorists"? I can certainly appreciate the value of being able to protect 6PB of data with only 9PB of storage capacity, instead of keeping two copies of 6PB each, the trade-off means that you will be accessing the majority of your data across your intranet, which could impact performance. But, if you are in an illicit or illegal business that could have a third of your facilities "seized by the government", then perhaps you shouldn't house your data centers there in the first place. Having two copies of 6PB each, in two "friendly nations", might make more sense.
(In reality, companies often keep way more than just two copies of data. It is not unheard of for companies to keep three to five copies scattered across two or three locations. Facebook keeps SIX copies of photographs you upload to their website.)
ChuckH argues that the governments that seize the three nodes won't have a complete copy of the data. However, merely having pieces of data is enough for governments to capture terrorists. Even if the striping is done at the smallest 512-byte block level, those 512 bytes of data might contain names, phone numbers, email addresses, credit cards or social security numbers. Hackers and computer forensics professionals take advantage of this.
You might ask yourself, "Why not just encrypt the data instead?" That brings me to the third trade-off, protection versus application performance. Over the past 30 years, companies had a choice, they could encrypt and decrypt the data as needed, using server CPU cycles, but this would slow down application processing. Every time you wanted to read or update a database record, more cycles would be consumed. This forced companies to be very selective on what data they encrypted, which columns or fields within a database, which email attachments, and other documents or spreadsheets.
An initial attempt to address this was to introduce an outboard appliance between the server and the storage device. For example, the server would write to the appliance with data in the clear, the appliance would encrypt the data, and pass it along to the tape drive. When retrieving data, the appliance would read the encrypted data from tape, decrypt it, and pass the data in the clear back to the server. However, this had the unintended consequences of using 2x to 3x more tape cartridges. Why? Because the encrypted data does not compress well, so tape drives with built-in compression capabilities would not be able to shrink down the data onto fewer tapes.
(I covered the importance of compressing data before encryption in my previous blog post
[Sock Sock Shoe Shoe].)
Like the trade-off between energy efficiency and abundant materials, IBM eliminated the trade-off by offering compression and encryption on the tape drive itself. This is standard 256-bit AES encryption implemented on a chip, able to process the data as it arrives at near line speed. So now, instead of having to choose between protecting your data or running your applications with acceptable performance, you can now do both, encrypt all of your data without having to be selective. This approach has been extended over to disk drives, so that disk systems like the IBM System Storage DS8000 and DS5000 can support full-disk-encryption [FDE] drives.
Certainly, something to think about!
technorati tags: , sunlight, solar cells, electricity, indium, gallium, cadmium, copper, tin, zinc, sulfur, selenium, thin+film, efficiency, EMC, Chuck Hollis, GeoProtect, Cleversafe, governement, seizure, Facebook, terrorists, encryption, forensics, hackers, protection, performance, disk, tape
Tonight PBS plans to air Season 38, Episode 6 of NOVA, titled [Smartest Machine On Earth]. Here is an excerpt from the station listing:
"What's so special about human intelligence and will scientists ever build a computer that rivals the flexibility and power of a human brain? In "Artificial Intelligence," NOVA takes viewers inside an IBM lab where a crack team has been working for nearly three years to perfect a machine that can answer any question. The scientists hope their machine will be able to beat expert contestants in one of the USA's most challenging TV quiz shows -- Jeopardy, which has entertained viewers for over four decades. "Artificial Intelligence" presents the exclusive inside story of how the IBM team developed the world's smartest computer from scratch. Now they're racing to finish it for a special Jeopardy airdate in February 2011. They've built an exact replica of the studio at its research lab near New York and invited past champions to compete against the machine, a big black box code -- named Watson after IBM's founder, Thomas J. Watson. But will Watson be able to beat out its human competition?"
Craig Rhinehart offers
[10 Things You Need to Know About the Technology Behind Watson].
An artist has come up with this clever
Dr. Jon Lenchner from IBM Research has a series of posts on
[How Watson "sees", "hears", and "speaks"] and [Selected Nuances].
Like most supercomputers, Watson runs the Linux operating system. The system runs 2,880 cores (90 IBM Power 750 servers, four sockets each, eight cores per socket) to achieve 80 [TeraFlops]. TeraFlops is the unit of measure for supercomputers, representing a trillion floating point operations. By comparison, Hans Morvec, principal research scientist at the Robotics Institute of Carnegie Mellon University (CMU) estimates that the [human brain is about 100 TeraFlops]. So, in the three seconds that Watson gets to calculate its response, it would have processed 240 trillion operations.
Several readers of my blog have asked for details on the storage aspects of Watson. Basically, it is a modified version of IBM Scale-Out NAS [SONAS] that IBM offers commercially, but running Linux on POWER instead of Linux-x86. System p expansion drawers of SAS 15K RPM 450GB drives, 12 drives each, are dual-connected to two storage nodes, for a total of 21.6TB of raw disk capacity. The storage nodes use IBM's General Parallel File System (GPFS) to provide clustered NFS access to the rest of the system. Each Power 750 has minimal internal storage mostly to hold the Linux operating system and programs.
When Watson is booted up, the 15TB of total RAM are loaded up, and thereafter the DeepQA processing is all done from memory. According to IBM Research, "The actual size of the data (analyzed and indexed text, knowledge bases, etc.) used for candidate answer generation and evidence evaluation is under 1TB." For performance reasons, various subsets of the data are replicated in RAM on different functional groups of cluster nodes. The entire system is self-contained, Watson is NOT going to the internet searching for answers.
On ZDnet, Steven J. Vaughan-Nichols welcomes our new [Linux Penguin Jeopardy overlords]. I have to say I share his enthusiasm!
technorati tags: IBM, Nova, Watson, #ibmwatson, Jeopardy, POWER7, p750, supercomputer, TeraFlops, disk, SONAS, GPFS, SAS, Craig Rhinehart, Jon Lenchner, Hans Morvec, Carnegie Mellon University, CMU
Wrapping up my coverage of the annual [2010 System Storage Technical University], I attended what might be perhaps the best session of the conference. Jim Nolting, IBM Semiconductor Manufacturing Engineer, presented the new IBM zEnterprise mainframe, "A New Dimension in Computing", under the Federal track.
The zEnterprises debunks the "one processor fits all" myth. For some I/O-intensive workloads, the mainframe continues to be the most cost-effective platform. However, there are other workloads where a memory-rich Intel or AMD x86 instance might be the best fit, and yet other workloads where the high number of parallel threads of reduced instruction set computing [RISC] such as IBM's POWER7 processor is more cost-effective. The IBM zEnterprise combines all three processor types into a single system, so that you can now run each workload on the processor that is optimized for that workload.
- IBM zEnterprise z196 Central Processing Complex (CPC)
Let's start with the new mainframe z196 central processing complex (CPC). Many thought this would be called the z11, but that didn't happen. Basically, the z196 machine has a maximum 96 cores versus z10's 64 core maximum, and each core runs 5.2GHz instead of z10's cores running at 4.7GHz. It is available in air-cooled and water-cooled models. The primary operating system that runs on this is called "z/OS", which when used with its integrated UNIX System Services subsystem, is fully UNIX-certified. The z196 server can also run z/VM, z/VSE, z/TPF and Linux on z, which is just Linux recompiled for the z/Architecture chip set. In my June 2008 post [Yes, Jon, there is a mainframe that can help replace 1500 servers], I mentioned the z10 mainframe had a top speed of nearly 30,000 MIPS (Million Instructions per Second). The new z196 machine can do 50,000 MIPS, a 60 percent increase!
(Update: Back in 2007, IBM and Sun mutually supported [OpenSolaris on an IBM System z mainframe]. Unfortunately, after Oracle acquired Sun, the OpenSolaris Governing Board has [grown uneasy over Oracle's silence] about the future of OpenSolaris on any platform. The OpenSolaris [download site] identifies 2009.06 as the latest release, but only for x86 and SPARC chip sets. Apparently, the 2010.03 release expected five months ago in March has slipped. Now it looks official that [OpenSolaris is Dead].)
The z196 runs a hypervisor called PR/SM that allows the box to be divided into dozens of logical partitions (LPAR), and the z/VM operating system can also act as a hypervisor running hundreds or thousands of guest OS images. Each core can be assigned a specialty engine "personality": GP for general processor, IFL for z/VM and Linux, zAAP for Java and XML processing, and zIIP for database, communications and remote disk mirroring. Like the z9 and z10, the z196 can attach to external disk and tape storage via ESCON, FICON or FCP protocols, and through NFS via 1GbE and 10GbE Ethernet.
- IBM zEnterprise BladeCenter Extension (zBX)
There is a new frame called the zBX that basically holds two IBM BladeCenter chassis, each capable of 14 blades, so total of 28 blades per zBX frame. For now, only select blade servers are supported inside, but IBM plans to expand this to include more as testing continues. The POWER-based blades can run native AIX, IBM's other UNIX operating system, and the x86-based blades can run Linux-x86 workloads, for example. Each of these blade servers can run a single OS natively, or run a hypervisor to have multiple guest OS images. IBM plans to look into running other POWER and x86-based operating systems in the future.
If you are already familiar with IBM's BladeCenter, then you can skip this paragraph. Basically, you have a chassis that holds 14 blades connected to a "mid-plane". On the back of the chassis, you have hot-swappable modules that snap into the other side of the mid-plane. There are modules for FCP, FCoE and Ethernet connectivity, which allows blades to talk to each other, as well as external storage. BladeCenter Management modules serve as both the service processor as well as the keyboard, video and mouse Local Console Manager (LCM). All of the IBM storage options available to IBM BladeCenter apply to zBX as well.
Besides general purpose blades, IBM will offer "accelerator" blades that will offload work from the z196. For example, let's say an OLAP-style query is issued via SQL to DB2 on z/OS. In the process of parsing the complicated query, it creates a Materialized Query Table (MQT) to temporarily hold some data. This MQT contains just the columnar data required, which can then be transferred to a set of blade servers known as the Smart Analytics Optimizer (SAO), then processes the request and sends the results back. The Smart Analytics Optimizer comes in various sizes, from small (7 blades) to extra large (56 blades, 28 in each of two zBX frames). A 14-blade configuration can hold about 1TB of compressed DB2 data in memory for processing.
- IBM zEnterprise Unified Resource Manager
You can have up to eight z196 machines and up to four zBX frames connected together into a monstrously large system. There are two internal networks. The Inter-ensemble data network (IEDN) is a 10GbE that connects all the OS images together, and can be further subdivided into separate virtual LANs (VLAN). The Inter-node management network (INMN) is a 1000 Mbps Base-T Ethernet that connects all the host servers together to be managed under a single pane of glass known as the Unified Resource Manager. It is based on IBM Systems Director.
By integrating service management, the Unified Resource Manager can handle Operations, Energy Management, Hypervisor Management, Virtual Server Lifecycle Management, Platform Performance Management, and Network Management, all from one place.
- IBM Rational Developer for System z Unit Test (RDz)
But what about developers and testers, such as those Independent Software Vendors (ISV) that produce mainframe software. How can IBM make their lives easier?
Phil Smith on z/Journal provides a history of [IBM Mainframe Emulation]. Back in 2007, three emulation options were in use in various shops:
- Open Mainframe, from Platform Solutions, Inc. (PSI)
- FLEX-ES, from Fundamental Software, Inc.
- Hercules, which is an open source package
None of these are viable options today. Nobody wanted to pay IBM for its Intellectual Property on the z/Architecture or license the use of the z/OS operating system. To fill the void, IBM put out an officially-supported emulation environment called IBM System z Professional Development Tool (zPDT) available to IBM employees, IBM Business Partners and ISVs that register through IBM Partnerworld. To help out developers and testers who work at clients that run mainframes, IBM now offers IBM Rational Developer for System z Unit Test, which is a modified version of zPDT that can run on a x86-based laptop or shared IBM System x server. Based on the open source [Eclipse IDE], the RDz emulates GP, IFL, zAAP and zIIP engines on a Linux-x86 base. A four-core x86 server can emulate a 3-engine mainframe.
With RDz, a developer can write code, compile and unit test all without consuming any mainframe MIPS. The interface is similar to Rational Application Developer (RAD), and so similar skills, tools and interfaces used to write Java, C/C++ and Fortran code can also be used for JCL, CICS, IMS, COBOL and PL/I on the mainframe. An IBM study ["Benchmarking IDE Efficiency"] found that developers using RDz were 30 percent more productive than using native z/OS ISPF. (I mention the use of RAD in my post [Three Things to do on the IBM Cloud]).
What does this all mean for the IT industry? First, the zEnterprise is perfectly positioned for [three-tier architecture] applications. A typical example could be a client-facing web-server on x86, talking to business logic running on POWER7, which in turn talks to database on z/OS in the z196 mainframe. Second, the zEnterprise is well-positioned for government agencies looking to modernize their operations and significantly reduce costs, corporations looking to consolidate data centers, and service providers looking to deploy public cloud offerings. Third, IBM storage is a great fit for the zEnterprise, with the IBM DS8000 series, XIV, SONAS and Information Archive accessible from both z196 and zBX servers.
To learn more, see the [12-page brochure] or review the collection of [IBM Redbooks]. Check out the [IBM Conferences schedule] for an event near you. Next year, the IBM Storage University will be held July 18-22, 2011 in Orlando, Flordia.
technorati tags: IBM, Technical University, zEnterprise, x86, POWER7, RISC, z/OS, Linux, AIX, OpenSolaris, Oracle, FICON, NFS, z196, zBX, DB2, SAO, IEDN, INMN, RDz, ISV, Eclipse, Cloud Computing