After my response to Jon Toigo on Drunken Data
withmy post [Elevenanswers about Deduplication
], Jon follows up with a request to validate the numbersquoted in the February 26 Press Release[IBM launches New “System z10” Mainframe
], particularly the estimate that a single mainframecan handle to the workload of 1500 x86 servers, in his post[A Bit More Blegging
]. The timing is perfect in that IBM launched[the next wave of Project Big Green
] today.To avoid sounding like an [editorial from the New York Sun
],I checked the facts, and spoke to the person in IBM who did all the calculations. Jon, as always, you have my permissionto publish this on your site if you want.
( I cannot take credit for coining the new term "bleg". I saw this term firstused over on the [FreakonomicsBlog]. If you have not yet read the book "Freakonomics", I highly recommend it! The authors' blog is excellent as well.)
For this comparison, it is important to figure out how much workload a mainframe can support, how much an x86 cansupport, and then divide one from the other. Sounds simple enough, right? And what workload should you choose?IBM chose a business-oriented "data-intensive" workload using Oracle database. (If you wanted instead a scientific"compute-intensive" workload, consider an [IBM supercomputer] instead, the most recent of which clocked in over 1 quadrillion floating point operations per second, or PetaFLOP.) IBM compares the following two systems:
- Sun Fire X2100 M2, model 1220 server (2-way)
IBM did not pick a wimpy machine to compare against. The model 1220 is the fastest in the series, with a 2.8Ghz x86-64 dual-core AMD Opteron processor, capable of running various levels of Solaris, Linux or Windows.In our case, we will use Oracle workloads running on Red Hat Enterprise Linux.All of the technical specifications are available at the[Sun Microsystems Sun Fire X1200] Web site.I am sure that there are comparable models from HP, Dell or even IBM that could have been used for this comparison.
- IBM z10 Enterprise Class mainframe model E64 (64-way)
This machine can run a variety of operating systems also, including Red Hat Enterprise Linux (RHEL). The E64 has four "multiple processor modules" called"processor books" for a total of 77 processing units: 64 central processors, 11 system assist processors (SAP) and 2 spares. That's right, spare processors, in case any others gobad, IBM has got your back. You can designate a central processor in a variety of flavors. For running z/VM and Linux operating systems, the central processors can be put into "Integrated Facility for Linux" (IFL) mode.On IT Jungle, Timothy Patrick Morgan explains the z10 EC in his article[IBM Launches 64-Way z10 Enterprise Class Mainframe Behemoth]. For more information on the z10 EC, see the 110-page [Technical Introduction], orread the specifications on the[IBM z10 EC] Web site.
Moving Oracle workloads from x86 over to mainframe is quite commonsince [IBM and Cisco joined forces to meet Linux On Mainframe demand]. For more information on consolidating x86 servers running Oracle over to a mainframe, read the [Quick Reference], IBMRedbook titled ["Using Oracle Solutionson Linux for System z"], or this [presentation by Jim Elliott, IBM System z Specialist].
In a shop full of x86 servers, there are production servers, test and development servers, quality assuranceservers, standby idle servers for high availability, and so on. On average, these are only 10 percent utilized.For example, consider the following mix of servers:
- 125 Production machines running 70 percent busy
- 125 Backup machines running idle ready for active failover in case a production machine fails
- 1250 machines for test, development and quality assurance, running at 5 percent average utilization
While [some might question, dispute or challenge thisten percent] estimate, it matches the logic used to justify VMware, XEN, Virtual Iron or other virtualization technologies. Running 10 to 20 "virtual servers" on a single physical x86 machine assumes a similar 5-10 percent utilization rate.
Note: The following paragraphs have been revised per comments received.
|Now the math. Jon, I want to make it clear I was not involved in writing the press release nor assisted with thesemath calculations. Please, don't shoot the messenger! Remember this cartoon where two scientists in white lab coats are writing mathcalculations on a chalkboard, and in the middle there is "and then a miracle happens..." to continue the rest ofthe calculations?|
In this case, the miracle is the number that compares one server hardware platform to another. I am not going to bore people with details like the number of concurrent processor threads or the differencesbetween L1 and L3 cache. IBM used sophisticated tools and third party involvement that I am not allowed to talk about, and I have discussed this post with lawyers representing
four (now five) different organizations already,so for the purposes of illustration and explanation only, I have reverse-engineered a new z10-to-Opteron conversion factor as 6.866 z10 EC MIPS per GHz of dual-core AMD Opteron for I/O-intensive workloads running only 10 percent average CPU utilization. Business applications that perform a lot of I/O don't use their CPU as much as other workloads.For compute-intensive or memory-intensive workloads, the conversion factor may be quite different, like 200 MIPS per GHz, as Jeff Savit from Sun Microsystems points out in the comments below.
Keep in mind that each processor is different, and we now have Intel, AMD, SPARC, PA-RISC and POWER (and others); 32-bit versus 64-bit; dual-core and quad-core; and different co-processor chip sets to worry about. AMD Opteron processors come in different speeds, but we are comparing against the 2.8GHz, so 1500 times 6.866 times 2.8 is 28,337. Since these would be running as Linux guestsunder z/VM, we add an additional 7 percent overhead or 2,019 MIPS. We then subtract 15 percent for "smoothing", whichis what happens when you consolidate workloads that have different peaks and valleys in workload, or 4,326 MIPS.The end is that we need a machine to do 26,530 MIPS. Thanks to advances in "Hypervisor" technological synergy between the z/VM operating system and the underlying z10 EC hardware, the mainframe can easily run 90 percent utilized when aggregating multiple workloads, so a 29,477 MIPS machine running at 90 percent utilization can handle these 26,530 MIPS.
N-way machines, from a little 2-way Sun Fire X2100 to the might 64-way z10 EC mainframe, are called "Symmetric Multiprocessors". All of the processors or cores are in play, but sometimes they have to taketurns, wait for exclusive access on a shared resource, such as cache or the bus. When your car is stopped at a red light, you are waiting for your turn to use the shared "intersection". As a result, you don't get linear improvement, but rather you get diminishing returns. This is known generically as the "SMP effect", and in IBM documentsthis as [Large System Performance Reference].While a 1-way z10 EC can handle 920 MIPS, the 64-way can only handle30,657 MIPS. The 29,477 MIPS needed for the Sun x2100 workload can be handled by a 61-way, giving you three extraprocessors to handle unexpected peaks in workload.
But are 1500 Linux guest images architecturally possible? A long time ago, David Boyes of[Sine Nomine Associates] ran 41,400 Linux guest images on a single mainframe using his [Test Plan Charlie], and IBM internallywas able to get 98,000 images, and in both cases these were on machines less powerful than the z10 EC. Neitherof these were tests ran I/O intensive workloads, but extreme limits are always worth testing. The 1500-to-1 reduction in IBM's press release is edge-of-the-envelope as well, so in production environments, several hundred guest images are probably more realistic, and still offer significant TCO savings.
The z10 EC can handle up to 60 LPARs, and each LPAR can run z/VM which acts much like VMware in allowing multipleLinux guests per z/VM instance. For 1500 Linux guests, you could have 25 guests each on 60 z/VM LPARs, or 250 guests on each of six z/VM LPARs, or 750 guests on two LPARs. with z/VM 5.3, each LPAR can support up to 256GB of memory and 32 processors, so you need at least two LPAR to use all 64 engines. Also, there are good reasons to have different guests under different z/VM LPARs, such as separating development/test from production workloads. If you had to re-IPLa specific z/VM LPAR, it could be done without impacting the workloads on other LPARs.
To access storage, IBM offers N-port ID Virtualization (NPIV). Without NPIV, two Linux guest images could not accessthe same LUN through the same FCP port because this would confuse the Host Bus Adapter (HBA), which IBM calls "FICON Express" cards. For example, Linux guest 1 asks to read LUN 587 block 32 and this is sent out a specific port, to a switch, to a disk system. Meanwhile, Linux guest 2 asks to read LUN 587 block 49. The data comes back to the z10 EC with the data, gives it to the correct z/VM LPAR, but then what? How does z/VM know which of the many Linux guests to give the data to? Both touched the same LUN, so it is unclear which made the request. To solve this, NPIV assigns a virtual "World Wide Port Name" (WWPN), up to 256 of them per physical port, so you can have up to 256 Linux guests sharing the same physical HBA port to access the same LUN.If you had 250 guests on each of six z/VM LPARs, and each LPAR had its own set of HBA ports, then all 1500 guestscould access the same LUN.
Yes, the z10 EC machines support Sysplex. The concept is confusing, but "Sysplex" in IBM terminology just means that you can have LPARs either on the same machine or on separate mainframes, all sharing the same time source, whether this be a "Sysplex Timer" or by using the "Server Time Protocol" (STP). The z10 EC can have STP over 6 Gbps Infiniband over distance. If you wantedto have all 1500 Linux guests time stamp data identically, all six z/VM LPARs need access to the shared time source. This can help in a re-do or roll-back situation for Oracle databases to complete or back-out "Units of Work" transactions. This time stamp is also used to form consistency groups in "z/OS Global Mirror", formerly called "XRC" for Extended Remote Distance Copy. Currently, the "timestamp" on I/O applies only to z/OS and Linux and not other operating systems. (The time stamp is done through the CDK driver on Linux, and contributed back to theopen source community so that it is available from both Novell SUSE and Red Hat distributions.)To have XRC have consistency between z/OS and Linux, the Linux guests would need to access native CKD volumes,rather than VM Minidisks or FCP-oriented LUNs.
Note: this is different than "Parallel Sysplex" which refers to having up to 32 z/OS images sharing a common "Coupling Facility" which acts as shared memory for applications. z/VM and Linux do not participate in"Parallel Sysplex".
As for the price, mainframes list for as little as "six figures" to as much as several million dollars, but I have no idea how much this particular model would cost. And, of course, this is just the hardware cost. I could not find the math for the $667 per server replacement you mentioned, so don't have details on that.You would need to purchase z/VM licenses, and possibly support contracts for Linux on System z to be fully comparable to all of the software license and support costs of the VMware, Solaris, Linux and/or Windows licenses you run on the x86 machines.
This is where a lot of the savings come from, as a lot of software is licensed "per processor" or "per core", and so software on 64 mainframe processors can be substantially less expensive than 1500 processors or 3000 cores.IBM does "eat its own cooking" in this case. IBM is consolidating 3900 one-application-each rack-mounted serversonto 30 mainframes, for a ratio of 130-to-1 and getting amazingly reduced TCO. The savings are in the followingareas:
- Hardware infrastructure. It's not just servers, but racks, PDUs, etc. It turns out to be less expensive to incrementally add more CPU and storage to an existing mainframe than to add or replace older rack-em-and-stack-emwith newer models of the same.
- Cables. Virtual servers can talk to each other in the same machine virtually, such as HiperSockets, eliminatingmany cables. NPIV allows many guests to share expensive cables to external devices.
- Networking ports. Both LAN and SAN networking gear can be greatly reduced because fewer ports are needed.
- Administration. We have Universities that can offer a guest image for every student without having a majorimpact to the sys-admins, as the students can do much of their administration remotely, without having physicalaccess to the machinery. Companies uses mainframe to host hundreds of virtual guests find reductions too!
- Connectivity. Consolidating distributed servers in many locations to a mainframe in one location allows youto reduce connections to the outside world. Instead of sixteen OC3 lines for sixteen different data centers, you could have one big OC48 line instead to a single data center.
- Software licenses. Licenses based on servers, cores or CPUs are reduced when you consolidate to the mainframe.
- Floorspace. Generally, floorspace is not in short supply in the USA, but in other areas it can be an issue.
- Power and Cooling. IBM has experienced significant reduction in power consumption and cooling requirementsin its own consolidation efforts.
All of the components of DFSMS (including DFP, DFHSM, DFDSS and DFRMM) were merged into a single product "DFSMS for z/OS" and is now an included element in the base z/OS operating system. As a result of these, customers typically have 80 to 90 percent utilization on their mainframe disk. For the 1500 Linux guests, however, most of the DFSMS features of z/OS do not apply. These functions were not "ported over" to z/VM nor Linux on any platform.
Note: DFSMS can backup or dump Linux on System z partitions or volumes. See this [Appendix C. HOWTO backup Linux data through z/OS] for details.
Instead, the DFSMS concepts have been re-implemented into a new product called "Scale-Out File Services" (SOFS) which would provide NAS interfaces to a blendeddisk-and-tape environment. The SOFS disk can be kept at 90 percent utilization because policies can place data, movedata and even expire files, just like DFSMS does for z/OS data sets. SOFS supports standard NAS protocols such as CIFS,NFS, FTP and HTTP, and these could be access from the 1500 Linux guests over an Ethernet Network Interface Card (NIC), which IBM calls "OSA Express" cards.
Lastly, IBM z10 EC is not emulating x86 or x86-64 interfaces for any of these workloads. No doubt IBM and AMD could collaborate together to come up with an AMD Opteron emulator for the S/390 chipset, and load Windows 2003 right on top of it, but that would just result in all kinds of emulation overhead.Instead, Linux on System z guests can run comparable workloads. There are many Linux applications that are functionally equivalent or the same as their Windows counterparts. If you run Oracle on Windows, you could runOracle on Linux. If you run MS Exchange on Windows, you could run Bynari on Linux and let all of your Outlook Expressusers not even know their Exchange server had been moved! Linux guest images can be application servers, web servers, database servers, network infrastructure servers, file servers, firewall, DNS, and so on. For nearly any business workload you can assign to an x86 server in a datacenter, there is likely an option for Linux on System z.
Hope this answers all of your questions, Jon. These were estimates based on basic assumptions. This is not to imply that IBM z10 EC and VMware are the only technologies that help in this area, you can certainly find virtualization on other systems and through other software.I have asked IBM to make public the "TCO framework" that sheds more light on this.As they say, "Your mileage may vary."
For more on this series, check out the following posts:
If in your travels, Jon, you run into someone interested to see how IBM could help consolidate rack-mounted servers over to a z10 EC mainframe, have them ask IBM for a "Scorpion study". That is the name of the assessment that evaluates a specific clientsituation, and can then recommend a more accurate estimate configuration.
technorati tags: Jon Toigo, DrunkenData, bleg, IBM, z10, EC, E64, mainframe, x86, AMD, Opteron, Sun, Fire, X2100, petaFLOP, Freakonomics, Red Hat, RHEL, IFL, VMware, Jim Elliott, Xen, Virtual Iron, Solaris, Linux, Windows, Project Big Green, Infiniband, STP, Sysplex, Scorpion study, MS Exchange, Bynari, Oracle
Continuing this week in Los Angeles, I went to some interesting sessions today at theSystems Technical Conference (STC08).
- System Storage Productivity Center (SSPC) - Install and Configuration
Dominic Pruitt, an IBM IT specialist in our Advanced Technical Support team, presented SSPC and howto install and configure it. For those confused between the difference of TotalStorage ProductivityCenter and System Storage Productivity Center, the former is pure software that you install on aWindows or Linux server, and the latter is an IBM server, pre-installed with Windows 2003, TotalStorageProductivity Center software, TPCTOOL command line interface, DB2 Universal Database, the DS8000 Element Manager, SVC GUI and CIMOM, and [PuTTY] rLogin/SSH/Telnet terminal application software.
SSPC speeds up the deployment of TotalStorage Productivity Center. The[SSPC Planning Worksheet] captures all of the pieces of information you need to activate the machine. On March 8,IBM simplified the [procedure to change the SSPC host name].
Of course, the problem with having a server pre-installed with a lot of software is that there is alwayssomeone that wants to customize it further. For those who just want to manage their DS8000 disk systems,for example, it is possible to uninstall the SVC GUI, CIMOM and PuTTY, and re-install them later when youchange your mind. As a general rule, it is not wise to mix CIMOMs on the same machine, as it might causeconflicts with TCP ports or Java level requirements, so if you want a different CIMOM than SVC, uninstallthe SVC CIMOM first. For those who have SVC, the SSPC replaces the SVC Master Console, so you can safelyturn off the SVC CIMOM on your existing SVC Master Consoles.
The base level is TotalStorage Productivity Center "Basic Edition", but you can upgrade the Productivity Centerfor Disk, Data and Fabric components with license keys. You can also run Productivity Center for Replication,but IBM recommends adding processor and memory to do this (IBM offers this as an orderable option).Whether you have the TotalStorage software or SSPC hardware, Productivity Center has a cool role-to-groups mapping feature.You can create user groups, either on the Windows server, the Active Directory, or other LDAP, and then map which roles should be assigned to users in each group.
Since Productivity Center manages a variety of different disk systems, it has made anattempt to standardize some terminology. The term "storage pool" refers to an extentpool on the DS8000, or a managed disk group on the SAN Volume Controller. Since the DS8000 can support both mainframe CKD volumes and LUNs for distributed systems, theterm "volume" refers to a CKD volume or LUN, and "disk" refers to the hard disk drive (HDD).
To help people learn Productivity Center, IBM offers single-day "remote workshops"that use Windows Remote Desktop to allow participants to install, customize and usethe software with no travel required.
- IBM Integrated Approach to Archiving
Dan Marshall, IBM global program manager for storage and data services on our Global Technology Services team, presented IBM's corporate-wide integration to support archive across systems, software and services.One attendee asked me why I was there, given that "archive" is one of my areas of subject matter expertise that I present often at the Tucson Executive Briefing Center. I find it useful to watch others present the material, even material that I helped to develop, to see a different slant or spin on each talking point.
Archive is one area that brings all parts of IBM together: systems, software and services.Dan provided a look at archive from the services angle, providing an objective unbiasedview of the different software and systems available to solve specific challenges.
- Encryption Key Manager (EKM) Design and Implementation
Jeff Ziehm, IBM tape technical sales specialist, presented IBM's EKM software, how it works in a tape environment, and how to deploy it in various environments. Since IBM is allabout being open and non-proprietary, the EKM software runs on Java on a variety ofIBM and non-IBM operating systems. IBM offers "keytool" command line interface (CLI) for the LTO4 and TS1120 tape systems, and "iKeyMan" graphical user interface (GUI) for theTS1120. Since it runs on Java, IBM Business Partners and technical support personneloften just [download and install EKM]onto their own laptops to learn how to use it.
- Virtual Tape Update
We had three presenters at this one. First, Jeff Mulliken, formerly from Diligent and now a full IBM employee, presented the current ProtecTier softwarewith the HyperFactor technology, then Abbe Woodcock, IBM tape systems, compared Diligent with IBM's TS7520 and just-announced TS7530virtual tape libraries, and finally Randy Fleenor, IBM tape sales leader, presented IBM's strategy going forward in tape virtualization.
Let's start with Diligent. The ProtecTier software runs on any x86-64 server withat least four cores and the correct Emulex host bus adapter (HBA) cards. Using Red HatEnterprise Linux (RHEL) as a base, the ProtecTier software performs its deduplication entirely in-lineat an "ingest rate" of 400-450 MB/sec. This is all possible using 4GB memory-resident "dictionary table" that can map up to 1 PB of back end physical storage, which could represent as much as 25PB of "nominal" storage. Theserver is then point-to-point or SAN-attached to Fibre Channel disk systems.
As we learned yesterday from Toby Marek's session, there are four ways to performdeduplication:
- full-file comparisons. Store only one copy of identical files.
- fixed-chunk comparisons. Files are carved up into fixed-size chunks, and each chunkis compared or hashed to existing chunks to eliminate duplicates.
- variable-chunk comparisons. Variable-length chunks are hashed or diffed to eliminate duplicate data.
- content-aware comparisons. If you knew data was in Powerpoint format, for example,you could compare text, photos or charts against other existing Powerpoint files toeliminate duplicates.
IBM System Storage N series Advanced Single Instance Storage (A-SIS) uses fixed-chunkmethod, and Diligent uses variable-chunk comparisons. Diligent does this using "dataprofiling". For example, let's say most of my photographs are pictures of people, buildings, landscapes, flowers and IT equipment. When I back these up, the Diligentserver "profiles" each, and determines if any existing data have a similar profilethat might have at least 50 percent similar content. Diligent than reads in the data that is mostly likely similar, does a byte-for-byte ["diff" comparison], and creates variable-lengthchunks that are either identical or unique to sections of the existing data. Theunique data is compressed with LZH and written to disk, and the sequential series of pointer segments representing the ingested file is written in a separate section on disk.
That Diligent can represent profiles for 1PB of data in as little as 4GB memory-residentdictionary is incredible. By comparison, 10TB data would require 10 million entries on a content-aware solution, and 1.25 billion entries for one based on hash-codes.
Abbe Woodcock presented the TS7530 tape system that IBM announced on Tuesday. It has some advantages over the current Diligent offering:
- Hardware-based compression (TS7520 and Diligent use software-based compression)
- 1200 MB/sec (faster ingest rate than Diligent)
- 1.7PB of SATA disk (more disk capacity than Diligent)
- Support for i5/OS (Diligent's emulation of ATL P3000 with DLT7000 tapes not supported on IBM's POWER systems running i5/OS)
- Ability to attach a real tape library
- NDMP backup to tape
- iSCSI attachment
- tape "shredding" (virtual equivalent of degaussing a physical tape to erase all previously stored data)
Randy Fleenor wrapped up the session telling us IBM's strategy going forward with all of thevirtual tape systems technologies. Until then, IBM is working on "recipes" or "bundles", puttingDiligent software with specific models of IBM System x servers and IBM System Storage DS4000 disk systemsto avoid the "do-it-yourself" problems of its current software-only packaging.
- Understanding Web 2.0 and Digital Archive Workloads
I got to present this in the last time slot of the day, just before everyone headed off to the [Westin Bonaventure hotel] for our big fancy barbecue dinner. Like my previous sessionon IBM Strategy, this session was more oriented toward a sales audience, but both garnereda huge turn-out and were well-received by the technical attendees.
This session was requested because these new applications and workloads are what is driving IBM to acquire small start-ups like XIV, deploy Scale-Out File Services (SOFS), and develop the innovative iDataPlex server rack.
The session was fun because it was a mix of explanation of the characteristics ofWeb 2.0 services; my own experience as a blogger and user of Google Docs, FlickR, Second Life andTivo; and an exploration in how database and digital archives will impact thegrowth in computing and storage requirements.
I'll expand on some of these topics in later blog posts.
technorati tags: IBM, SSPC, System Storage, TotalStorage, Productivity Center, Windows 2003, Linux, TPCTOOL, DB2, DS8000, SVC, GUI, CLI, CIMOM, PuTTY, SSH, Java, LDAP, CKD, Volume, LUN, HDD, workshops, Dominic Pruitt, Dan Marshall, Global Technology Services, GTS, archive, Jeff Ziehm, EKM, keytool, iKeyMan, LTO4, TS1120, VTL, Diligent, ProtecTier, HyperFactor, Jeff Mulliken, Abbe Woodcock, Randy Fleenor, TS7520, TS7530, x86-64, RHEL, deduplication, A-SIS, diff, LZH, compression, i5/OS, DLT7000, P3000, NDMP, iSCSI, shredding, DS4000, Westin Bonaventure, Web2.0, digital archive, XIV, SOFS, iDataPlex, Google Docs, Second Life, FlickR, Tivo
This is the second day of our Systems Technology Conference (STC08) in Los Angeles, California.We have over 700 break-out sessions, packed in 16 times slots across 47 rooms.
- IBM Strategy - the New Enterprise Data Center
In February, IBM launched its corporate-wide strategy for the "new enterprise data center", which I discussedalready in my post[Is your data center ready for the future?]
My session was the first in the morning, at 8:30am, but managed to pack the room full of people. A few looklike they just rolled in from Brocade's special get-together in Casey's Irish Pub the night before.I presented how IBM's storage strategy for the information infrastructure fits into the greater corporate-wide themes.To liven things up, I gave out copies of my book[Inside System Storage: Volume I] to those who asked or answered the toughest questions.
- Data Deduplication and IBM Tivoli Storage Manager (TSM)
IBM Toby Marek compared and contrasted the various data deduplication technologies and products available, andhow to deploy them as the repository for TSM workloads. She is a software engineer for our TSM software product,and gave a fair comparison between IBM System Storage N series Advanced Single Instance Storage (A-SIS), IBMDiligent, and other solutions out in the marketplace.If you are going to combine technologies, then it isbest to dedupe first, then compress, and finally encrypt the data. She also explained about the many cleverways that TSM does data reduction at the client side greatly reduces the bandwidth traffic over the LAN,as well as reducing disk and tape resources for storage. This includes progressive "incremental forever" backup for file selection, incremental backups for databases, and adaptive sub-file backup.Because of these data reduction techniques, you may not get as much benefit as deduplication vendors claim.
- The Business Value of Energy Efficiency Data Centers
Scott Barielle did a great job presenting the issues related to the Green IT data center. He is part of IBM"STG Lab Services" team that does energy efficiency studies for customers. It is not unusual for his teamto find potential savings of up to 80 percent of the Watts consumed in a client's data center.
IBM has done a lot to make its products more energy efficient. For example, in the United States, most datacenters are supplied three-phase 480V AC current, but this is often stepped down to 208V or 110V with powerdistribution units (PDUs). IBM's equipment allows for direct connection to this 480V, eliminating the step-downloss. This is available for the IBM System z mainframe, the IBM System Storage DS8000disk system, and larger full-frame models of our POWER-based servers, and will probably be rolled out to someof our other offerings later this year. The end result saves 8 to 14 percent in energy costs.
(Last October, IBM Randy Malik made a similar case in his presentation:[HighVoltage Distributions])
Scott had some interesting statistics. Typical US data centers only spend about 9 percent of their IT budgeton power and cooling costs. The majority of clients that engage IBM for an energy efficiency study are not tryingto reduce their operational expenditures (OPEX), but have run out, or close to running out, of total kW ratingof their current facility, and have been turned down by their upper management to spend the average $20 million USDneeded to build a new one. The cost of electricity in the USA has risen very slowly over the past 35 years, andis more tied the to fluctuations of Natural Gas than it is to Oil prices.(a recent article in the Dallas News confirmed this:["As electricity rates go up, natural gas' high prices, deregulation blamed"])
- Cognos v8 - Delivering Operational Business Intelligence (BI) on Mainframe
Mike Biere, author of the book [BusinessIntelligence for the Enterprise], presented Cognos v8 and how it is being deployed for the IBMSystem z mainframe. Typically, customers do their BI processing on distributed systems, but 70 percent of the world's business data is on mainframes, so it makes sense to do yourBI there as well. Cognos v8 runs on Linux for System z, connecting to z/OS via [Hypersockets].
There are a variety of other BI applications on the mainframe already, including DataQuant,AlphaBlox, IBI WebFocus and SAS Enterprise Business Intelligence. In addition to accessing traditional onlinetransaction processing (OLTP) repositories like DB2, IMS and VSAM, using the [IBM WebSphere ClassicFederation Server], Cognos v8 can also read Lotus databases.
Business Intelligence is traditionally query, reporting and online analytics process (OLAP) for the top 10 to 15 percent of the company, mostly executives andanalysts, for activities like business planning, budgeting and forecasting. Cognos PowerPlay stores numericaldata in an [OLAP cube] for faster processing.OLAP cubes are typically constructed with a batch cycle, using either "Extract, Transfer, Load" [ETL], or "Change Data Capture" [CDC], which playsto the strength of IBM System z mainframe batch processing capabilities.If you are not familiar with OLAP, Nigel Pendse has an article[What is OLAP?] for background information.
Over the past five years, BI is now being more andmore deployed for the rest of the company, knowledge workers tasked with doing day-to-day operations. Thisphenomenom is being called "Operational" Business Intelligence.
For more on this, see the IBM Systems Magazine article [Upgrade Your Mainframe with Operational Business Intelligence].
- IBM GPFS - Fundamentals and What's New
IBM Glen Corneau, who is on the Advanced Technical Support team for AIX and System p, presented the IBMGeneral Parellel File System (GPFS), which is available for AIX, Linux-x86 and Linux on POWER.Unfortunately, many of the questions were related to Scale Out File Services (SOFS), which my colleague GlennHechler was presenting in another room during this same time slot.
GPFS is now in its 11th release since its introducing in 1997. All of the IBM supercomputers on the [Top 500 list] use GPFS. The largest deployment of GPFS is 2241 nodes.A GPFS environment can support up to 256 file systems, each file system can have up to 2 billion filesacross 2 PB of storage. GPFS supports "Direct I/O" making it a great candidate for Oracle RAC deployments.Oracle 10g automatically detects if it is using GPFS, and sets the appropriate DIO bits in the stream totake advantage of GPFS features.
Glen also covered the many new features of GPFS, such as the ability to place data on different tiers ofstorage, with policies to move to lower tiers of storage, or delete after a certain time period, all conceptswe call Information Lifecycle Management. GPFS also supports access across multiple locations and offersa variety of choices for disaster recovery (DR) data replication.
Perhaps the only problem with conferences like this is that it can be an overwhelming["fire hose"] of information!
technorati tags: IBM, STC08, new enterprise data center, storage, strategy, deduplication, TSM, Diligent, A-SIS, Green IT, Toby Marek, Scott Barielle, , Brocade, High Voltage, PDU, DS8000, disk, systems, mainframe, POWER, Randy Malik, OPEX, Natural Gas, Electricity, BI, Cognos, OLAP, GPFS, Glen Corneau, AIX, Top500, Disaster Recovery
This week I'm in Los Angeles for the Systems Technology Conference (STC '08).We have over 1900 IT professionals attending, of which 1200 IBMers from North America, Latin America,and Asia Pacific regions, as well as another 350 IBM Business Partners. The rest, including me, are world wideor from other areas.
Last January, IBM reorganized its team to be more client-focused. Instead of focused on products, we are nowclient-centric, and have teams to cover our large enterprise systems through direct sales force, business systemsfor sales through our channel business partners, and industry systems for specific areas like deep computing,digital surveillance and retail systems solutions.
In addition to 788 sessions to attend these next four days, we had a few main tent sessions.My third line (my boss' boss' boss) David Gelardi presented Enterprise Systems. This is the group I am in.
Akemi Watanabe presented for Business Systems. Her native language is Japanese, so to do an entire talk inEnglish was quite impressive. Her focus is on SMB accounts, those customers with less than 1000 employeesthat are looking for easy-to-use solutions. She mentioned IBM's new [Blue Business Platform] which includesLotus Foundation Start, an Application Integration Toolkit, and the Global Application Marketplace.
Part of this process is the merger of System p and System i into "POWER" systems, and then offering both midrangeand enterprise versions of these that run AIX, i5/OS and Linux on POWER. It turns out that only 9 percent of ourSystem i customers are only on this platform. Another 87 percent have Windows, so it makes sense to offer i5/OSon BladeCenter, to consolidate Windows servers from HP, Dell or Sun over to IBM.
Meanwhile, IBM's strategy to support Linux has proven successful. 25 percent of x86 servers now run Linux. IBMhas 600 full-time developers for Linux, over 500 of which contributed to the latest 2.6 kernel development. Our ["chiphopper"] program has successfullyported over 900 applications. There are now over 6500 applications that run on Linux applications, on our strategic alliances with Red Hat (RHEL) and Novell (SUSE) distributions of Linux.
Her recommendation to SMB reps: learn POWER systems, BladeCenter, and Linux. I agree!
Mary Coucher presented Industry systems. In addition to the game chips for the Sony Playstation, Nintendo Wii,and Microsoft Xbox-360, this segment focuses on Digital Video Surveillance (DVS), Retail Solutions, Healthcare and Life sciences (HCLS), OEM and embedded solutions, and Deep computing. She mentioned our recently announcediDataPlex solution.
IBM is focused on "real-world-aware" applications, which includes traffic, crime, surveillance, fraud, andRFID enablement. These are streams of data that happen real-time, that need to be dealt with now, not later.
Most people know that IBM has the majority of the top 500 supercomputers, but few may not realize that IBMalso has delivered solutions to the top 100 green companies. IBM success is explained in more detail in this[Press Release].
The group split up to four different platform meetings: Storage, Modular, Power, and Mainframe. Barry Rudolphpresented for the Storage platform. He talked about the explosion in information, business opportunities,risk and cost management. IBM has shifted from being product-focused, to the stack of servers and storage,to our latest focus on solutions across the infrastructure. He mentioned our DARPA win for [PERCS] which stands for productive,easy-to-use, reliable computing system.
Exciting times at IBM!
technorati tags: STC08, IBM, North America, Latin America, Asia Pacific, David Gelardi, enterprise systems, Akemi Watanabe, Mary Coucher, HCLS, real-world aware, OEM, deep computing, DVS, RHEL, SUSE, HP, Sun, Dell, Storage, Modular, Power, Mainframe, Barry Rudolph, DARPA, PERCS
My theme this week was to focus on "Do-it-Yourself" solutions, such as the "open storage" concept presentedby Sun Microsystems, but it has morphed into a discussion on vendor lock-in. Both deserve a bit of furtherexploration.
There were several reasons offered on why someone might pursue a "Do-it-Yourself" course of action.
- Building up skills
In my post [Simply Dinners and Open Storage], I suggested that building a server-as-storage solution based on Sun's OpenSolaris operating system could serve to learn more about [OpenSolaris], and by extension, the Solaris operating system.Like Linux, OpenSolaris is open source and has distributions that run on a variety of chipsets, from Sun's ownSPARC, to commodity x86 and x86-64 hardware. And as I mentioned in my post [Getting off the island], a version of OpenSolaris was even shown to run successfully on the IBM System z mainframe.
"Learning by Doing" is a strong part of the [Constructivism] movement in education. TheOne Laptop Per Child [OLPC] uses this approach. IBM volunteers in Tucson and 40other sites [help young students build robots]constructed from [Lego Mindstorms]building blocks.Edward De Bono uses the term [operacy] to refer to the"skills of doing", preferred over just "knowing" facts and figures.
However, I feel OpenSolaris is late to the game. Linux, Windows and MacOS are all well-established x86-based operating systems that most home office/small office users would be familiar with, and OpenSolaris is positioning itself as "the fourth choice".
In my post[WashingtonGets e-Discovery Wakeup Call], I suggested that the primary motivation for the White House to switch from Lotus Notes over to Microsoft Outlookwas familiarity with Microsoft's offerings. Unfortunately, that also meant abandoning a fully-operational automated email archive system, fora manual do-it-yourself approach copying PST files from journal folders.
Familiarity also explains why other government employees might print out their emails and archive them on paperin filing cabinets. They are familiar with this process, it allows them to treat email in the same manner as they have treated paper documents in the past.
- Cost, Control and Unique Requirements
The last category of reasons can often result if what you want is smaller or bigger than what is availablecommercially. There are minimum entry-points for many vendors. If you want something so small that it is notprofitable, you may end up doing it yourself. On the other end of the scale, both Yahoo and Google ended up building their data centers with a do-it-yourself approach, because no commercial solutions were available atthe time. (IBM now offers [iDataPlex], so that has changed!)
While you could hire a vendor to build a customized solution to meet your unique requirements, it might turn outto be less costly to do-it-yourself. This might also provide some added control over the technologies and components employed. However, as EMC blogger Chuck Hollis correctly pointed out for[Do-it-yourself storage],your solution may not be less costly than existingoff-the-shelf solutions from existing storage vendors, when you factor in scalability and support costs.
Of course, this all assumes that storage admins building the do-it-yourself storage have enough spare time to do so. When was the last time your storage admins had spare time of any kind?Will your storage admins provide the 24x7 support you could get from established storage vendors? Will theybe able to fix the problem fast enough to keep your business running?
From this, I would gather that if you have storage admins more familiar with Solaris than Linux, Windows or MacOS,and select commodity x86 servers from IBM, Sun, HP, or Dell, they could build a solution that has less vendor lock-in than something off-the-shelf from Sun. Let's explore the fears of vendor lock-in further.
- The storage vendor goes out of business
Sun has not been doing so well, so perhaps "open storage" was a way to warn existing Sun storage customers thatbuilding your own may be the next alternative.The New York Times title of their article says it all:["Sun Microsystems Posts Loss and Plans to Reduce Jobs"]. Sun is a big company, so I don't expect them to close their doors entirely this year,but certainly fear of being locked-in to any storage vendor's solution gets worse if you fear the vendor might go out of business.
- The storage vendor will get acquired by a vendor you don't like
We've seen this before. You don't like vendor A, so you buy kit from vendor B, only to have vendor A acquire vendorB after your purchase. Surprise!
- The storage vendor will not support new applications, operating systems, or other new equipment
Here the fear is that the decisions you make today might prevent you from choices you want to make in the future.You might want to upgrade to the latest level of your operating system, but your storage vendor doesn't supportit yet. Or maybe you want to upgrade your SAN to a faster bandwidth speed, like 8 Gbps, but your storage vendordoesn't support it yet. Or perhaps that change would require re-writing lots of scripts using the existingcommand line interfaces (CLI). Or perhaps your admins would require new training for the new configuration.
- The storage vendor will raise prices or charge you more than you expect on follow-on upgrades
For most monolithic storage arrays, adding additional disk capacity means buying it from the same vendor as the controller. I heard of one company recently who tried to order entry-level disk expansion drawer, at a lower price, solely to move the individual disk drives into a higher-end disk system. Guess what? It didn't work. Most storage vendors would not support such mixed configurations.
If you are going to purchase additional storage capacity to an existing disk system, it should cost no more thanthe capacity price rate of your original purchase. IBM offers upgrades at the going market rate, but not all competitors are this nice. Some take advantage of the vendor lock-in, charging more for upgrades and pocketing the difference as profit.
Vendor lock-in represents the obstacles in switching vendors in the event the vendor goes out of business, failsto support new software or hardware in the data center, or charges more than you are comfortable with. These obstacles can make it difficult to switch storage vendors, upgrade your applications, or meet otherbusiness obligations. IBM SANVolume Controller and TotalStorage Productivity Center can help reduce or eliminate many of these concerns. IBMGlobal Services can help you, as much or as little, as you want in this transformation. Here are the four levelsof the do-it-yourself continuum:
|Let me figure it out myself||Tell me what to do||Help me do it||Do it for me|
|This is the self-service approach. Go to our website, download an [IBM Redbook], figure out whatyou need, and order the parts to do-it-yourself.||IBM Global Business Services can help understand your business requirementsand tell you what you need to meet them.||IBM Global Technology Services can help design, assemble and deploy asolution, working with your staff to ensure skill and knowledge transfer.||IBM Managed Storage Services can manage your storage, on-site at your location, or at an IBM facility. IBM provides a varietyof cloud computing and managed hosting services.|
So, if you are currently a Sun server or storage customer concerned about these latest Sun announcements, give IBM a call, we'll help you switch over!
technorati tags: do-it-yourself, OpenSolaris, Solaris, Sun, Linux, SPARC, Yahoo, Google, iDataPlex, x86, x86-64, x64, mainframe, EMC, Chuck Hollis, HP, Dell, SAN, monolithic, disk, storage, system, arrays, open storage, NYT, New York Times, vendor lock-in, IBM, Global Services, GBS, GTS, SVC, TotalStorage, Productivity Center, Managed Storage Services, cloud computing