“In times of universal deceit, telling the truth will be a revolutionary act.”
-- George Orwell
Well, it has been over two years since I first covered IBM's acquisition of the XIV company. Amazingly, I still see a lot of misperceptions out in the blogosphere, especially those regarding double drive failures for the XIV storage system. Despite various attempts to [explain XIV resiliency] and to [dispel the rumors], there are still competitors making stuff up, putting fear, uncertainty and doubt into the minds of prospective XIV clients.
Clients love the IBM XIV storage system! In this economy, companies are not stupid. Before buying any enterprise-class disk system, they ask the tough questions, run evaluation tests, and all the other due diligence often referred to as "kicking the tires". Here is what some IBM clients have said about their XIV systems:
“3-5 minutes vs. 8-10 hours rebuild time...”
-- satisfied XIV client
“...we tested an entire module failure - all data is re-distributed in under 6 hours...only 3-5% performance degradation during rebuild...”
-- excited XIV client
“Not only did XIV meet our expectations, it greatly exceeded them...”
-- delighted XIV client
In this blog post, I hope to set the record straight. It is not my intent to embarrass anyone in particular, so instead will focus on a fact-based approach.
- Fact: IBM has sold THOUSANDS of XIV systems
XIV is "proven" technology with thousands of XIV systems in company data centers. And by systems, I mean full disk systems with 6 to 15 modules in a single rack, twelve drives per module. That equates to hundreds of thousands of disk drives in production TODAY, comparable to the number of disk drives studied by [Google], and [Carnegie Mellon University] that I discussed in my blog post [Fleet Cars and Skin Cells].
- Fact: To date, no customer has lost data as a result of a Double Drive Failure on XIV storage system
This has always been true, both when XIV was a stand-alone company and since the IBM acquisition two years ago. When examining the resilience of an array to any single or multiple component failures, it's important to understand the architecture and the design of the system and not assume all systems are alike. At it's core, XIV is a grid-based storage system. IBM XIV does not use traditional RAID-5 or RAID-10 method, but instead data is distributed across loosely connected data modules which act as independent building blocks. XIV divides each LUN into 1MB "chunks", and stores two copies of each chunk on separate drives in separate modules. We call this "RAID-X".
Spreading all the data across many drives is not unique to XIV. Many disk systems, including EMC CLARiiON-based V-Max, HP EVA, and Hitachi Data Systems (HDS) USP-V, allow customers to get XIV-like performance by spreading LUNs across multiple RAID ranks. This is known in the industry as "wide-striping". Some vendors use the terms "metavolumes" or "extent pools" to refer to their implementations of wide-striping. Clients have coined their own phrases, such as "stripes across stripes", "plaid stripes", or "RAID 500". It is highly unlikely that an XIV will experience a double drive failure that ultimately requires recovery of files or LUNs, and is substantially less vulnerable to data loss than an EVA, USP-V or V-Max configured in RAID-5. Fellow blogger Keith Stevenson (IBM) compared XIV's RAID-X design to other forms of RAID in his post [RAID in the 21st Centure].
- Fact: IBM XIV is designed to minimize the likelihood and impact of a double drive failure
The independent failure of two drives is a rare occurrence. More data has been lost from hash collisions on EMC Centera than from double drive failures on XIV, and hash collisions are also very rare. While the published worst-case time to re-protect from a 1TB drive failure for a fully-configured XIV is 30 minutes, field experience shows XIV regaining full redundancy on average in 12 minutes. That is 40 times less likely than a typical 8-10 hour window for a RAID-5 configuration.
A lot of bad things can happen in those 8-10 hours of traditional RAID rebuild. Performance can be seriously degraded. Other components may be affected, as they share cache, connected to the same backplane or bus, or co-dependent in some other manner. An engineer supporting the customer onsite during a RAID-5 rebuild might pull the wrong drive, thereby causing a double drive failure they were hoping to avoid. Having IBM XIV rebuild in only a few minutes addresses this "human factor".
In his post [XIV drive management], fellow blogger Jim Kelly (IBM) covers a variety of reasons why storage admins feel double drive failures are more than just random chance. XIV avoids load stress normally associated with traditional RAID rebuild by evenly spreading out the workload across all drives. This is known in the industry as "wear-leveling". When the first drive fails, the recovery is spread across the remaining 179 drives, so that each drive only processes about 1 percent of the data. The [Ultrastar A7K1000] 1TB SATA disk drives that IBM uses from HGST have specified 1.2 million hours mean-time-between-failures [MTBF] would average about one drive failing every nine months in a 180-drive XIV system. However, field experience shows that an XIV system will experience, on average, one drive failure per 13 months, comparable to what companies experience with more robust Fibre Channel drives. That's innovative XIV wear-leveling at work!
- Fact: In the highly unlikely event that a DDF were to occur, you will have full read/write access to nearly all of your data on the XIV, all but a few GB.
Even though it has NEVER happened in the field, some clients and prospects are curious what a double drive failure on an XIV would look like. First, a critical alert message would be sent to both the client and IBM, and a "union list" is generated, identifying all the chunks in common. The worst case on a 15-module XIV fully loaded with 79TB data is approximately 9000 chunks, or 9GB of data. The remaining 78.991 TB of unaffected data are fully accessible for read or write. Any I/O requests for the chunks in the "union list" will have no response yet, so there is no way for host applications to access outdated information or cause any corruption.
(One blogger compared losing data on the XIV to drilling a hole through the phone book. Mathematically, the drill bit would be only 1/16th of an inch, or 1.60 millimeters for you folks outside the USA. Enough to knock out perhaps one character from a name or phone number on each page. If you have ever seen an actor in the movies look up a phone number in a telephone booth then yank out a page from the phone book, the XIV equivalent would be cutting out 1/8th of a page from an 1100 page phone book. In both cases, all of the rest of the unaffected information is full accessible, and it is easy to identify which information is missing.)
If the second drive failed several minutes after the first drive, the process for full redundancy is already well under way. This means the union list is considerably shorter or completely empty, and substantially fewer chunks are impacted. Contrast this with RAID-5, where being 99 percent complete on the rebuild when the second drive fails is just as catastrophic as having both drives fail simultaneously.
- Fact: After a DDF event, the files on these few GB can be identified for recovery.
Once IBM receives notification of a critical event, an IBM engineer immediately connects to the XIV using remote service support method. There is no need to send someone physically onsite, the repair actions can be done remotely. The IBM engineer has tools from HGST to recover, in most cases, all of the data.
Any "union" chunk that the HGST tools are unable to recover will be set to "media error" mode. The IBM engineer can provide the client a list of the XIV LUNs and LBAs that are on the "media error" list. From this list, the client can determine which hosts these LUNs are attached to, and run file scan utility to the file systems that these LUNs represent. Files that get a media error during this scan will be listed as needing recovery. A chunk could contain several small files, or the chunk could be just part of a large file. To minimize time, the scans and recoveries can all be prioritized and performed in parallel across host systems zoned to these LUNs.
As with any file or volume recovery, keep in mind that these might be part of a larger consistency group, and that your recovery procedures should make sense for the applications involved. In any case, you are probably going to be up-and-running in less time with XIV than recovery from a RAID-5 double failure would take, and certainly nowhere near "beyond repair" that other vendors might have you believe.
- Fact: This does not mean you can eliminate all Disaster Recovery planning!
To put this in perspective, you are more likely to lose XIV data from an earthquake, hurricane, fire or flood than from a double drive failure. As with any unlikely disaster, it is best to have a disaster recovery plan than to hope it never happens. All disk systems that sit on a single datacenter floor are vulnerable to such disasters.
For mission-critical applications, IBM recommends using disk mirroring capability. IBM XIV storage system offers synchronous and asynchronous mirroring natively, both included at no additional charge.
For more about IBM XIV reliability, read this whitepaper [IBM XIV© Storage System: Reliability Reinvented]. To find out why so many clients LOVE their XIV, contact your local IBM storage sales rep or IBM Business Partner.
technorati tags: IBM, XIV, DDF, RAID-5, RAID-10, RAID-X, RAID-6, RAID-DP, HP, EVA, HDS, USP-V, EMC, CLARiiON, V-Max, Disaster Recovery, HGST, UltraStar, A7K1000
Well, it's Tuesday, and that means IBM announcements!
IBM kicks EMC in the teeth with the announcement of System Storage Easy Tier, a new feature available at no additional charge
on the DS8700 with the R5.1 level microcode. Barry Whyte introduces the concept in his [post this morning
]. I will use SLAM (sub-LUN automatic movement) to refer generically to IBM Easy Tier and EMC FAST v2. EMC has yet to deliver FAST v2, and given that they just recently got full-LUN FAST v1 working a few months ago, it might be next year
before you see EMC sub-LUN FAST v2.
Here are the key features of Easy Tier on the DS8700:
- Sub-LUN Automatic Movement
IBM made it really easy to implement this on the DS8700. Today, you have "extent pools" that can be either SSD-only or HDD-only. With this new announcement, we introduce "mixed" SSD+HDD extent pools. The hottest extents are moved to SSD, and cooler extents are moved down to HDD. The support applies to both Fixed block architecture (FBA) LUNs as well as Count-Key-Data (CKD) volumes. In other words, an individual LUN or CKD volume can have some of its 1GB extents on SSD and other extents on FC or SATA disk.
- Entire-LUN Manual Relocation
Entire-LUN Manual Relocation (ELMR, pronounced "Elmer"?) is similar to what EMC offers now with FAST v1. With this feature, you can now relocate an entire LUN non-disruptively from any extent pool to any other extent pool. You can relocate LUNs from an SSD-only or HDD-only pool over to a new Easy Tier-managed "mixed" pool, or take a LUN out of Easy Tier management by moving it to an SSD-only or HDD-only pool. Of course, this support also applies to both Fixed block architecture (FBA) LUNs as well as Count-Key-Data (CKD) volumes.
This feature also can be used to relocate LUNs and CKD volumes from FC to SATA pools, from RAID-10 to RAID-5 pools, and so on.
- Pool Mergers
What if you already have SSD-only and HDD-only pools and want to use Easy Tier? You can now merge pools to create a "mixed" pool.
- SSD Mini-Packs
Before this announcement, you had to buy 16 solid-state drives at a time, called Mega-packs. Now, you can choose to buy just 8 SSD at a time, called Mini-packs. It turns out that just moving as little as 10 percent of your data from Fibre Channel disk over to Solid-State with Easy Tier can result in up to 300 to 400 percent performance improvement. IBM plans to publish formal SPC-1 benchmark results using Easy Tier-managed mixed extent pool in a few weeks.
- Storage Tier Advisor Tool (STAT)
Don't have SSD yet, or not sure how awesome Easy Tier will be for your data center? The IBM Storage Tier Advisor Tool will analyze your extents and estimate how much benefit you will derive if you implement Easy Tier with various amounts of SSD. Those clients with R5.1 microcode on their DS8700 can download from the [DS8700 FTP site].
To learn more, see the [Easy Tier landing page] and the 10-page Easy Tier chapter in [DS8000 Introduction and Planning Guide]. IBM also had announcements regarding LTO-5 tape, N series and XIV storage systems, which I will get to in later posts.
technorati tags: IBM, Easy Tier, SLAM, ELMR, DS8700, SSD, HDD, extent pool, FBA, CKD, LUN, FC, SATA, disk, storage, RAID-5, RAID-10, Mega-Pack, Mini-Pack
My colleagues, Harley Puckett (left) and Jack Arnold (right) were highlighted in today's Arizona Daily Star, our local newspaper, as part of an article on IBM's success and leadership in the IT storage industry. At 1400 employees here in Tucson, IBM is Southern Arizona's 36th largest employer.
Highlighted in the article:
- DS8700 with the new Easy Tier feature
- TS7650 ProtecTIER virtual tape library with data deduplication capability
- LTO-5 tape and the new Long Term File System (LTFS)
- XIV with the new 2TB drive, for a maximum per-rack usable capacity of 161 TB.
Read the full article [IBMers Crank Out 4 New Offerings To Handle Data Deluge]
technorati tags: , Arizona Daily Star, IBM Tucson, DS8700, Easy Tier, ProtecTIER, Deduplication, LTO-5, LTFS, XIV, IBM, Tucson, Arizona
Continuing my discussion of this week's announcements of IBM storage products, I will cover the announcements that double storage capacity per footprint.
- Linear Tape Open - Generation 5
IBM announced [LTO-5 drives], the TS2250 half-height and the TS2350 full-height drives, as well as support for LTO-5 drives in its various tape libraries: TS3100, TS3200, and TS3500. The native 1.5TB capacity of the LTO-5 cartridge is nearly double the 800GB capacity of the LTO-4 predecessor. With 2:1 compression, that's 3TB of data per cartridge! Performance-wise, the data transfer rate is 140 MB/sec, about 17 percent improvement over the 120MB/sec of the LTO-4 technology. The TS2250, TS2350, TS3100 and TS3200 now all offer dual-SAS ports for higher availability.
LTO-5 carries forward many of the advancements of past generations. For example, LTO-5 continues the G-2/G-1 "backward compatibility" architecture, which means that the LTO-5 drive can read LTO-3 and LTO-4 cartridges, and can write LTO-4 cartridges. Like the LTO-3 and LTO-4, the same LTO-5 drive can read and write WORM or regular rewriteable cartridges. Like the LTO-4, the LTO-5 offers drive-level data-at-rest encryption. These use a symmetric 256-bit AES key, managed by IBM Tivoli Key Lifecycle Manager (TKLM).
One thing that is new in LTO-5 is the Long Term File System [LTFS] available on the TS2250 and TS2350, which allows you to treat the tape as a hierarchical file system, with files and folders, that you can drag and drop like any other file system.
- XIV storage system
IBM [doubles the capacity of the XIV storage system] by supporting 2TB SATA drives. A full 15-module frame can hold up to 161TB of usable capacity. The smallest 6-module system with 2TB can hold up to 55TB of usable capacity. At this time, all of the drives in an XIV must be the same type, so we do not yet allow intermix of 1TB and 2TB in the same frame. The 2TB are more energy efficient, with a full 15-module frame consuming on average 6.7 kVA, compared to 7.8 kVA for the 1TB drives. The performance is roughly the same, so if, for example, your application workload got 3700 IOPS per module with 1TB drives, it will get about the same 3700 IOPS per module with 2TB drives.
- TS7650 ProtecTIER Data Deduplication
IBM now supports [many-to-one virtual tape volume mirroring] on the ProtecTIER. In other words, you can have two or more locations sending data to a single ProtecTIER disaster recovery site.
- N series disk system
The EXN1000 and EXN3000 can now double in capacity with 2TB SATA drives. These can be attached to the N3000 entry-level models, such as the N3400.
- DS3000 disk system
The DS3200, DS3300 and DS3400, as well as their related expansion drawers, now supports 2TB SATA drives. This means that a single control unit with three expansion drawers can hold up to 96TB of raw capacity (48 drives).
- DS8700 disk system
The DS8700 also now supports 2TB SATA drives, for a maximum raw capacity over 2PB, as well as new 600GB Fibre Channel drives. Now that IBM offers [Easy Tier] functionality, pairing Solid State Drives with slower, energy-efficient SATA disk makes a lot of financial sense.
That's a lot of announcements! As always, feel free to dig into each of the links to learn more about each product.
technorati tags: IBM, LTO-5, TS2250, TS2350, TS3100, TS3200, TS500, AES, TKLM, LTFS, XIV, 2TB, TS7650, TS7650G, EXN1000, EXN3000, N3400, DS3200, DS3300, DS3400, DS8700, SATA
They say "Great Minds think alike" and that imitation is "the sincerest form of flattery." Both of these quotes came to mind when I read fellow blogger Chuck Hollis' (EMC) excellent April 7th blog post [The 10 Big Ideas That Are Shaping IT Infrastructure Today]. Not surprisingly, some of his thoughts are similar to those I had presented two weeks ago in my March 22nd post [Cloud Computing for Accountants]. Here are two charts that caught my eye:
- Telephone Operators
On page 13 of my deck, I had an old black and white photo of telephone operators, as part of a section on the history of selecting "cloud" as the iconic graphic to represent all networks. Chuck has this same graphic on his chart titled "#1 The Industrialization of IT Infrastructure".
Looks like Chuck and I use the same "stock photo" search facility!
- Arms Dealers
On page 45 on my deck, I had a list of major "arms dealers" that deliver the hardware and software components needed to build Cloud Computing. Chuck has a similar chart, titled "#2 The Consolidation of the IT Industry", but with some interesting differences.
Let's look at some of the key differences:
- The left-to-right order is slightly different. I chose a 1-2-4-2-1 symmetrical pattern purely on aesthetic reasons. My presentation was to a bunch of accountants, and so I was trying not to make it sound like an "Infomercial" for IBM products and offerings. My sequence is roughly chronological, in that Oracle announced its intention to acquire Sun, then Cisco, VMware and EMC announced their VCE coalition, followed closely by Cisco, VMware and NetApp announcing they work together well also, followed by [HP extended alliance with Microsoft] on Jan 13, 2010. As the IT marketplace is maturing, more and more customers are looking for an IBM-like one-stop shopping experience, and certainly various "mini-mall" alliances have formed to try to compete in this space.
- I had HP and Microsoft in the same column, referring only to the above-mentioned January announcement. HP is all about private cloud hardware infrastructures, but Microsoft is all about "three screens and the public cloud", so not sure how well this alliance will work out from a Cloud Computing perspective. This was not to imply that the other stacks don't work well with Microsoft software. They all do. Perhaps to avoid that controversy, Chuck chose to highlight HP's acquisition of EDS services instead.
- I used the vendor logos in their actual colors. Notice that the colors black, blue and red occur most often. These happen to be the three most popular ballpoint pen ink colors found on the very same paper documents these computer companies are trying to eliminate. Paper-less office, anyone? Chuck chose instead to colorize each stack with his own color scheme. While blue for IBM and orange for Sun Microsystems make some sense, it is not clear if he chose green for Cisco/VMware/EMC for any particular reason. Perhaps he was trying to subtly imply that the VCE stack is more energy efficient? Or maybe the green refers to money to indicate that the VCE stack is the most expensive? Either way, I would pit IBM's server/storage/software stack up against anything of comparable price from these other stacks in any energy efficiency bake-off.
- What about the Cisco/VMware/NetApp combination? All three got together to assure customers this was a viable combination. IBM is the number one reseller of VMware, and VMware runs great with IBM's N series NAS storage, so I do not dispute Cisco's motivation here. It makes sense for Cisco to two-time EMC in this manner. Why should Cisco limit itself to a single storage supplier? Et tu VMware? Having VMware chose NetApp over its parent company EMC was a bit of a shock. No surprise that Chuck left NetApp out of his chart.
- No love for Dell? I give Dell credit for their work with Virtual Desktop Images (VDI), and for embracing Ubuntu Linux for their servers. Dell's acquisitions of EqualLogic iSCSI-based disk systems and Perot Systems for services are also worth noting. Dell used to resell some of EMC's gear, but perhaps that relationship continues to fade away, as I [predicted back in 2007]. Chuck's decision to leave Dell off his chart speaks volumes to where this relationship stands, and where it is going.
Perhaps we are all in just one big ["echo chamber"], as we are all coming up with similar observations, talking to similar customers, and reviewing similar market analyst reports. I am glad, at least this time, that Chuck and I for the most part agree where the marketplace is going. We live in interesting times!
technorati tags: , Chuck Hollis, EMC, IBM, HP, Microsoft, telephone operators, IT infrastructure, cloud computing, Cisco, VMware, , , Oracle, Sun, NetApp, EqualLogic, iSCSI, NAS, Dell, VDI, Ubuntu, Linux