- Resilient storage requires worthwhile testing tools
- Smart array may signal storage shift
- FCoE catches fire at SNW
- Infiniband: Back from the dead again?
- Thin provisioning, fat savings
- We have a distinguished guest...
- It's already debate over FCoE
- A crowd of bloggers gathering at SNW
- A Vista on storage - restore woes
- A Vista on storage
June 20, 2008 | Comments: (0)
Resilient storage requires worthwhile testing tools
Taking the latest technologies for a test spin isn't my only reward as a reviewer. Learning how to use slick testing tools is a thrill all its own.
Sometimes those two overlap -- which was the case recently when I put Cisco's new Nexus 5000 to the test using technology from Ixia, a vendor that specializes in IP testing.
You may be wondering what storage has to do with IP testing. Well, let me set the stage.
Cisco's Nexus 5000 is a revolutionary switch that implements Fibre Channel over Ethernet. FCoE, as I've discussed before, shepherds Ethernet to never drop a frame and to speak FC. In doing so, it provides customers an easy-to-deploy consolidation path that allows them to preserve their FC investments. Moreover, it gives vendors the opportunity to sell CNAs (converged network adapters).
If you are not familiar with CNAs, think of a host adapter card that mounts a FC ASIC, GbE/10GbE ASIC, and a FCoE ASIC and you're close to the truth. Mount a CNA or two to your server, link them to GbE ports on the Nexus 5000, and your application can jump to a connected fabric and access all the SAN volumes you give it permission to. This may sound complicated, but it isn't, as I found in my test.
You'll have to wait for the published results of my evaluation, but my first impression is that the Nexus 5000, along with the Emulex and Qlogic CNAs I used in my test bed, delivers handsomely on the promise of FCoE.
What I can say is that the Nexus 5000 is a well-rounded Ethernet switch that can host a number of FC connections side by side with numerous 10G connections. And that's were Ixia's tools -- in particular, its IP Performance Tester -- came into play. These tools proved invaluable in testing the switch under strain.
For example, as part of my test, I measured how fast the Nexus 5000 could learn thousands of MAC addresses. The only way to run that test would have been to suddenly connect the switch to thousands of Ethernet devices. But because putting that many machines in one test location is nearly impossible, I had to find an alternative. Ixia did the trick.
A few clicks in Ixia's easy-to-use GUI and I had set up a script to simulate connections with 16,000 MAC addresses. Better yet, I was able to test the switch's responsiveness by pushing addresses at it at different intervals to see how well it kept up. Thanks to the IP Performance Test, you'll have the results of that evaluation soon.
IP Performance Tester offers many other worthwhile testing features, but I would like to mention one in particular. Imagine the equivalent of an Excel spreadsheet that Ixia's tool refreshes dynamically to give you insight into, say, the traffic levels your switches' ports are experiencing. Add to that the ability to create your own formulas, using the content of the cells provided by IxExplorer, and if that doesn't excite you, not much else will.
Do you have a large IP network? What tools are you using for testing and troubleshooting? Leave a comment below or e-mail me.
Posted by Mario Apicella on June 20, 2008 03:00 AM
April 18, 2008 | Comments: (0)
Smart array may signal storage shift
Although you often find statements that seem critical of disk drives in this column, I never criticize disk drives per se, but rather how we use them. As of late, I've been interested in the innovative use of disk drives in storage arrays. The latest news from Xiotech fits right into this discussion, providing further proof that the storage industry is fast turning evolution into revolution in rethinking its approach to disk drives.
[ For recent disk drive commentary, see "Upgrading to solid state," "Time to bury big-drive 'diskosaurs'," and "Diskosaurs bite back." ]
As you may have read, Xiotech began shipping two new storage solutions last week, the Emprise 5000 and the Emprise 7000. The arrays differ in capacity and scalability, but what's interesting is that neither mounts the usual crowd of disk drives we're used to seeing in every storage solution. Instead, they replace the drives with ISE (Intelligent Storage Elements) datapacs, sealed containers that offer 10 to 20 times the capacity of a single disk drive plus a bunch of built-in reliability features that will make your head spin.
The technology was brought on board in a little-publicized acquisition from Seagate, which included the transfer of about 100 employees to Xiotech.
What's in those containers? Depending on requirements, an ISE datapac can host 3.5-inch disk drives, 2.5-inch drives, or even SSDs (solid-state drives) while maintaining the same form factor. Judging from a presentation I saw, two datapacs can comfortably fit in a 3U enclosure, leaving room for controllers, power supply, battery modules, and the like.
As for the "intelligence" part, using smart controllers and advanced data protection techniques, ISE can stretch the MTBF (mean time between failure) to hundreds or thousands of times that of a single disk drive. Moreover, when a drive fails, the ISE controller will attempt a factory rebuild of the drive (think very low-level formatting), which should bring the device back to life more than 50 percent of the time, according to Xiotech. And when the drive is repaired, data rebuilding occurs automatically and without disruption.
Compare this with the hours or days it can take to rebuild the content of a large SATA drive, for example, and you understand why Xiotech has invested in delivering this technology to its customers.
But how does the performance of these new ISE arrays compare with traditional approaches to disk drives in competing arrays? Because some storage vendors (EMC, for example) snub independently reviewed benchmarks such as those published at the SPC (Storage Performance Council) site, comparisons are difficult. But there are many other vendors that publish their own benchmark results, including Fujitsu, HP, IBM, and Sun. Xiotech has also filed SPC-1 (random I/O) and SPC-2 (throughput) benchmark results for review, and it seems that the Emprise line has fared well, better than its competitors, not necessarily in sheer performance, but in price/performance ratios.
Stats and announcements aside, I can't help but think that there is more behind this move and that some questions have not yet been asked, let alone answered.
For example, Seagate has been toying with this technology in secret for some time and must have presented it to other major storage vendors as well. Is it possible that no other vendor found what would eventually become ISE interesting at the time Seagate was working on it? Hard to believe.
Other questions bug me even more. Seagate had to have pushed the technology as a possible alternative to selling individual drives, so why weren't there any other takers in the storage industry? Are storage vendors so enamored of their own data reliability improvement techniques that they refuse all outside alternatives? Were vendors afraid that Seagate would sell the datapacs and eat away at their competitive advantage? Was Seagate forced (or gently advised) not to open that Pandora's box?
We may never get a satisfactory answer to those questions, but Xiotech answered a clear yes, albeit after a moment of silence, when I asked whether it would offer the technology to other vendors.
There you have it: ISE can be also yours; it's only a matter of price.
Posted by Mario Apicella on April 18, 2008 03:00 AM
April 11, 2008 | Comments: (0)
Storage Networking World came and went this week, and judging from the hype, FCoE (Fibre Channel over Ethernet) is coming out of the conference with a head full of steam.
Understandably so, I may add, because this technology, which only a year ago was little more than a jot on a napkin, is poised to make its mark on the storage market.
Of course, running FC transport over Ethernet has never been impossible from a technical perspective. Politics, as this rather animated discussion proves, have played a large role in ensuring the two technologies have grown up in isolation. Fast-forward 12 months to this recent StorageMojo post and you'll see that the debate over FCoE hasn't slowed down a bit. And it's not surprising, given how much is at stake for both customers and vendors.
Obviously, some iSCSI vendors don't like the idea of FC invading the Ethernet space with a protocol that promises to be much more effective than alternative technologies, such as iFCP. After all, this would nullify, or at least weaken, their traditional advantages -- lower-cost connectivity to servers, for example.
By using a unified transport protocol, customers can access any storage volume from any server, assuming they have the proper gear. Vendors such as Emulex and QLogic are already shipping CNAs (converged network adapters) -- cards that, in essence, connect to an Ethernet wire but speak FC, allowing them to communicate with FC storage arrays.
There is, however, a gap in this picture. At some point, the FC protocol must jump from the Ethernet wire to the storage fabric. And here's where switch vendors come into the picture. On Tuesday, Cisco announced the availability of the Nexus 5000, a new switch especially designed for FCoE. We'll see how the switch measures up when I get a chance to review it, but to give an indication of the stakes, Cisco anticipates $14 billion over the next five years in FCoE-related revenue. Cisco partners will secure $4 billion of that succulent pie, the estimate goes.
Whether that mark is feasible depends entirely on how deeply customers buy into the promise of FCoE. Two areas where FCoE will make the most sense are server and datacenter consolidation. In these instances, the savings of running FC over Ethernet will likely far outweigh the costs. Other reasons for buying into FCoE -- to simplify wiring at the server, for example, or to extend the reach of the storage fabric across the WAN -- would require case-by-case analysis before committing to the technology.
Emulex, QLogic, VMware, and Dell are among the many vendors partnering with Cisco on FCoE. Intel is also jumping on this promising new market and will soon deliver FCoE adapters and chips integrated on server motherboards.
Of course, despite all this buzz, there is no reason to dismiss iSCSI. For customers who do not have a large investment in FC storage, iSCSI is still a very viable and less expensive proposition. More importantly, some so-called iSCSI solutions offer benefits that go beyond cheap and easy-to-use transport between servers and storage. Think, for example, of Intransa or Isilon: Their multinode, resilient architectures are unrivaled in the FC space. I would argue that customers buy those solutions because of their architectures, with iSCSI being just the icing on the cake.
Nevertheless, it would be naive to ignore that FCoE changes the rules of the storage game. Once the initial dust settles and we get used to the fact that both protocols can run over 10-Gig Ethernet, customers will be less distracted by debates over which transport is better and will instead focus on the storage solution that better serves their business. This alone should make the new protocol a welcome addition to the storage discussion.
Posted by Mario Apicella on April 11, 2008 03:00 AM
March 14, 2008 | Comments: (0)
Infiniband: Back from the dead again?
The need for I/O virtualization and lack of practical short-term alternatives breathes new life into Infiniband
For many IT managers, Infiniband immediately brings HPC (high-performance computing) to mind. In fact, this much-snubbed connectivity protocol has found vindication in HPC, where its minimal latency, sustained transfer rate, and gentle impact on CPU cycles have found much appreciation.
It's not that Infiniband is unknown among the lower tiers of enterprise computing, but vendors have predominantly favored other protocols, such as FC (Fibre Channel) and, more recently, the controversial newcomer, iSCSI. Which is a pity, because Infiniband could very well be the ultimate server connectivity protocol -- at least for the time being.
Because of this, expect Infiniband to soon undergo a renaissance of sorts, one which -- if memory serves -- will not be the first and -- if history has any bearing on the future -- will likely be as short-lived as previous rebirths.
This dated but well-conceived white paper from Cisco will provide a worthwhile refresher on Infiniband, for those who might need it. I'll quote a short excerpt from that document:
From a protocol perspective, the InfiniBand architecture consists of four layers: physical, logical, network, and transport. These layers are analogous to Layers 1 through 4 of the OSI protocol stack. Missing Presentation, application and session layer.
The protocol high-level layers mentioned above never got formalized, which could be the reason for, or perhaps the effect of, its limited diffusion in corporate infrastructure. In fact, for many IT managers who don't administer HPC applications, Infiniband is just an unnecessary nuisance, one that doesn't add much to what FC can already do.
If this provocative statement describes you, server consolidation is unlikely a significant issue at your shop. Think instead of having 20 to 30 VMs on one server: If, say, six of those apps require a dedicated FC pipe, you will have to install six FC HBAs. Double that for dual-path redundancy.
Under these circumstances, expect to see a spaghetti dish of cables spew from your server -- an unsightly mess that is as difficult to manage as it is expensive. Can you scale? Perhaps, but only up to a point; sooner or later you will run out of slots for your HBAs. Got some iSCSI targets to reach? Then add some GbE NICs to that server and as many Ethernet cables.
How does Infiniband help if you find yourself in that predicament? A good solution comes from Xsigo (pronounced seego) Systems, a new company that launched its first product, I/O Director in September.
A look at the photo shows that I/O Director hosts 24 Infiniband ports on the top row. The lower part of the appliance can load FC or GbE cards. Install one or, for redundancy, two Infiniband HCAs (host channel adapters) on each physical server and connect them to the I/O Director. The lower ports then connect to the switch or to the storage targets using FC or GbE.
The first advantage of deploying I/O Director is to reduce both cable clutter and the number of adapters at the server. Each HCA can dish out as much as 20Gbps, which can support multiple FC or iSCSI connections. But my favorite aspect of the Xsigo solution is the software that runs on the appliance.
Via your browser or a VMware plug-in (a CLI is also available), you can create virtual HBAs and virtual NICs and map them to the appropriate storage target. Xsigo claims it is the only vendor to offer this capability, and to the best of my knowledge, that assertion is correct.
Imagine the flexibility a virtual adapter will bring to your datacenter, in essence allowing you to provide the proper connection to newly created VMs on demand. The guest machine cannot tell the difference, viewing the virtual adapter as if it were physical. For example, during a demo I saw, a new hardware wizard popped up immediately after a virtual adapter was added to a Windows guest machine.
You can also move virtual adapters from one VM to another, even across different physical hosts: Think a poor man's VMotion, as Xsigo aptly put it during my demo. In fact, the receiving machine will also inherit all storage volumes reached by that adapter -- an easy way to move application data quickly to a new environment when something goes wrong.
Will Xsigo and other Infiniband-centered solutions bring the protocol back to the corporate deployment realm? My guess is yes, but only for as long as its current bandwidth advantage will continue. 100-Gigabit Ethernet could be the death sentence for Infiniband, but I don't see that coming on too fast.
Have you already deployed Infiniband? Care to quickly share your use cases and experience?
Posted by Mario Apicella on March 14, 2008 03:00 AM
October 02, 2007 | Comments: (0)
Thin provisioning, fat savings
I'm planning to take a look at thin provisioning. Seems plenty of vendors are talking up the technology but I am trying to gauge where the trade-offs lie.
Reads one of my messages this week. Please scroll down or click for more.
One article I read likened it to the airline industries approach to overselling flights in the anticipation that some people won't show up.In the case of the thin provisioning, the writer says, there's a risk that a storage admin may run out of storage due to incorrect forecasting of storage needs. Is that a flaw in the technology, or does it just mean that companies using thin provisioning need to plan better?
Any insights you can provide would be very helpful.
Well, thin provisioning is not a new topic nor uncharted territory, but I'll be glad to give you an overview. You may also want to check other blogs, for example Tony Asaro's at the Enteprise Strategy Group.
That analogy with airline reservations fits well thin provisioning, but only up to a point.
A dirty little secret of the storage world is that most storage requirements are grossly overestimated, often by an order of magnitude or more. It's not that storage admins are incompetent but often someone else, for example the database admin or the developer, calls the shots on how much capacity a project needs.
So you have a typical situation where there is authority ("I'll tell you how much storage I need") without responsibility ("it's not my job to optimize the amount of storage we buy or run").
In addition, estimating the capacity requirements for a new project is not easy and people tend to overestimate to be on the safe side.
Case three: Taking back excess storage after the new project is in production and the requirements are finally clear, is rarely done, so like a diamond, an overallocation of storage is forever.
Thin provisioning gives the storage admin the ability to provision the application server with the amount of space requested, say 100 GB, but taking away from the storage pool only a fraction, say 20 GB, if that's the amount "thin provisioned".
The objective of thin provision is not to forecast future use correctly, but to stretch the current capacity and cover requirements totaling a much larger amount of storage.
In many cases, the thinly provisioned amount, plus adjustments for normal data growth, gives enough capacity throughout the life of the application. Nevertheless, each company should find out what's their real data growth rate and add more capacity according to that trend.
If they don't do those adjustments correctly, at some point one of the applications will run out of space, which is never a pretty thing to live through.
However, I don't buy that incorrect forecasting is a risk associated only with thin provisioning. Storage capacity needs to be closely monitored regardless if doing fat or thin provisioning. If an admin runs out of space with thin provisioning, he or she would have botched also with fat provisioning, IMO.
Anyway, when done correctly thin provisioning is a big money saver because a purchase or a lease can be delayed until really needed, and will cover only the increment suggested by the growth trend. For a small company thin provisioning may not make much of a difference, because the granularity of their purchases can't go below a certain level ( you can't buy a third of a storage enclosure, for example).
For larger companies that buy millions of dollars of storage, thin provisioning is too good an opportunity to pass on because in addition to the financial aspects (later disbursement e and so on) they also save kilowatts/hr or at least push back in time the demand.
Hope this helps, but you know where to find me if it doesn't...
Posted by Mario Apicella on October 2, 2007 01:37 PM
April 24, 2007 | Comments: (0)
We have a distinguished guest...
... Julian Satran who just posted a comment to the FCoE debate that you should not miss.
Julian doesn't usually need an introduction, but in case you have been away :>) here is a short bio. Hope it's still accurate.
Posted by Mario Apicella on April 24, 2007 12:27 PM
April 13, 2007 | Comments: (0)
It's already debate over FCoE
It will take probably years before FCoE becomes available, but there are already some very strong opinions from FC and iSCSI vendors.
Read for example what Jeffrey Nick, Senior Vice President and Chief Technology Officer at EMC has to say:
Converging Fibre Channel protocols with an Ethernet infrastructure while retaining fibre channel's enterprise-class quality of service and management features will result in a major win on many levels for both endusers and IT vendors.
No surprises there, FCoE promises a smother ride over different media without leaving the FC transport. So EMC and other vendors with a large portfolio of FC products just love it.
In sharp contrast is what Eric Schott Director of Product Management at EqualLogic has to say on FCoE:
With the proposed standard Fibre Channel vendors have basically admitted the inevitable domination of Ethernet. This is a strategic retreat, with a goal of holding onto their existing investment in the FC architecture for as long as possible. They've now acknowledged Fibre Channel as the wire cannot compete with 10Gig Ethernet.
Maybe so, but what about FC as a transport protocol? Mr. Schott has a strong view on that too:
It's funny, with this news Fibre Channel vendors are planning to get rid of the FC wire, but are keeping the protocol, which is hard for customers to implement and manage. This is this first shoe to drop.
I can't wait to read what's the other shoe:
Fibre Channel customers will have to absorb a lot of costs to convert their FC infrastructure to Ethernet as a transport, so why don't the vendors have the customers move to iSCSI protocol to simplify their storage networks? The answer of course is they would cause the other shoe to drop, the admission that IP is excellent for storage, just as it is for all other enterprise applications. This admission is tantamount to admitting FC needs to move aside for iSCSI. Customers should examine very carefully this half step to FCoE, and consider taking the full step to iSCSI themselves
There you have it. The vision of a single wire, Ethernet, and a single transport, iSCSI, can be appealing but does anyone really believe that customers will swarm away from FC?
I don't. But will be interesting to compare FC and iSCSI performance over the same 10G wire, when FCoE products will ship.
Posted by Mario Apicella on April 13, 2007 06:41 AM
April 06, 2007 | Comments: (0)
A crowd of bloggers gathering at SNW
Only two years ago you could have counted storage bloggers without taking your socks off and now we can put together a small crowd.
Anil Gupta is trying to persuade that crowd (or at least part of it) to meet face to face at SNW in San Diego.
If you navigate the blogsphere and are going to SNW check out Anil Gupta blog and let him know if you want to meet with other bloggers while you are there.
I won't be there this time, but there is always a next time, right?
Posted by Mario Apicella on April 6, 2007 07:30 AM
February 08, 2007 | Comments: (0)
A Vista on storage - restore woes
For many people the next logical action after installing Vista will be trying to restore some of their files from a backup taken on Windows XP.
Agreed? Unfortunately this apparently simple task is not as easy as it should be when you are on Vista.
Let me start saying that Vista has a new backup and restore application. Why? If you really want to know more, read this PowerPoint presentation by Dan Stevenson, Lead Program Manager for Microsoft.
Impressive, isn't it? It's too bad that the new backup apps doesn't support tapes, but you can still use, USB drives, flash drives, DVDs, CDs and network shares. Who cares about tapes.
Too bad number two: The new backup apps in Vista is not backward compatible with the old one that shipped with Windows XP. In plain English, to restore files from a backup made on XP you need to install a restore utility on Vista, yessir! As it happens, that restore utility looks like and walks like the old XP backup restore duck, but I am getting ahead of myself.
So, no big deal, let's install the restore utility on Vista. To make things easier the Vista restore window has a link to the download page for the restore utility on Microsoft. Cool!.
That's what my to do list for restoring those files has become:
1) go to the download page
2) validate your version as a genuine Microsoft Windows
3) download restore utility
I go past those three points like a hot knife through butter, but only to be stopped by a warning message saying that I need to install or turn on Removable Storage Management to run the restore utility. Instructions to do that are on the Backup Restore Utility Web page. Do I want to go there?
Well I answered yes but only to find out that there are no instructions on that page, not on how to activate Removable Storage Management anyway. Moreover, nowhere in the process tells me where to find the new utility after the install.
Luckily, you can find all what you need here, and in one page.
So there is a happy ending to my restore story after all, but not thanks to Microsoft.
Lesson learned: If you plan to restore files in Vista from an old Windows backup be prepared to a not so smooth sailing. Stock on patience and allow plenty of time.
To be continued.
Posted by Mario Apicella on February 8, 2007 09:52 AM
February 07, 2007 | Comments: (0)
What should you expect to happen to your personal storage gear if you move to Vista?
I wasn't anticipating much of change from Windows XP and haven't found anything that could be a show-stopper so far. However I ran into a few interesting things in this my first and unsupervised (not too mention unbiased by any previous training) encounter with Vista.
I installed Vista Ultimate on a new SATA drive just added to my SFF HP Compaq D530, a machine chosen for no other reason than it was the only one I could safely spare without borrowing or stealing.
Not much to tell about the install, but despite 1 additional GB of RAM, which brings the total to 1.5 GB, Vista is still poo-pooing that machine with a WEI (Windows Experience Index) rating of 1.0, the lowest you can get.
Anyway, Vista is still responsive and doesn't lag too much, actually I am posting this from the Vista PC.
I digress. Back to storage, a "simple" installation of the OS set my 60GB hard-drive back 9.5 GB. After adding FlashGet, FireFox and a few extensions I am getting close to 14GB, and haven't even touched Office.
My point is: If you are serious about backups better forget the idea of using DVDs (worse yet CDs) and get some gadget with serious capacity. So far Vista has recognized (and worked with) any USB drive I've attached to it, including an Iomega REV 70, home of my first full backup.
Obviously, I have many more devices to test on Vista, but I found a locked door, at least for now, with the Storage Central 101 that doesn't have drivers for Vista yet. I am keeping an hopeful eye on the NetGear site.
Back to the grind now, but feel free to post with snippets from your experience with Vista, or with comments and suggestions.
To be continued.
Posted by Mario Apicella on February 7, 2007 09:52 AM
November 09, 2006 | Comments: (0)
Dell jumps on the RDX train too
"Looks like Prostor is doing something right" - was my opening line for a conversation with Steve Georgis, CEO of ProStor Systems. "We certainly hope so" - Georgis joked in response.
Of course Mr. Georgis and the management team at ProStor know very well what they are doing right, which is defining the technical framework for what promises to be the most serious contender to tape drives for small and medium business.
"Our strategy at ProStor has been to really create the first true multi vendor standard for removable disks backup and archive products. ProStor is the company doing the design and we are licensing others to do the manufacturing"- Georgis explains.
The RDX (removable disk extendable) standard defines the characteristics of the cartridge, built around a 2.5" SATA drive, and of its drive bay.
The cartridge includes a layer of shock absorbing material that makes the drive able to survive a fall from 1 meter (a little more than 3 feet), even over a hard surface like concrete.
That protective layer shields the cartridge also from electrostatic discharge, but should a cartridge become damaged, users will immediately notice because an embedded alarm led, usually green, will turn red after the initial diagnostic cycle.
Today the RDX cartridge is available in 40GB, 80GB and 120GB, but future larger capacity cartridges will mount seamlessly in the same drive bay. Also interesting is that cartridges from different vendors are compatible, which is an undeniable advantage over single branded solutions such as Quantum GoVault and Iomega REV
With a promised transfer rate of about 30MB/sec, higher for larger drives, SMB customers should get backup and restore speeds that they couldn't probably afford on tape drives.
The drive bay has also some interesting characteristics, with software than monitors the health of the cartridge and a 1 million hours MTBF. Prostor estimates a 13 years lifespan at full duty cycle, and a 10 years shelf life for RDX.
Tandberg Data, one of the first adopters, have been shipping products for some time. In fact, I have one of their RDX drives in my lab, but I haven't dropped the cartridge yet to prove how it can survive a fall.
About one month ago Imation became also interested in RDX and should begin shipping their first drives soon.
Which brings me finally to yesterday's announcement from Dell that will soon begin shipping an external USB version and an internal SATA version of RDX.
Is RDX the beginning of the end for the tormented entry-level tape drives segment?
Possibly, but mounting one cartridge at a time may not be fun if a backup is larger than the drive capacity. I wouldn't be surprised if an RDX autoloader will get to market next year.
Posted by Mario Apicella on November 9, 2006 12:38 PM
October 30, 2006 | Comments: (0)
Betting on the future of storage
Predictive market isn't exactly a new technology, but as far as I can tell, it was never applied to the storage industry -- until now.
What's predictive market? It's a market research tool that employs the stock market metaphor -- securities, traders, and shares -- to prognosticate the future outcome of an idea. Traders who register with a predictive market are assigned a certain amount of virtual money to finance their transactions, usually at no charge for them.
It works like this: If you like an idea, you buy its shares just as you would do with actual securities. Obviously, if many people buy shares of the same idea the price will go up, which tells an observer that the idea has some merit. Put two competing ideas on the market, and traders will soon decide which one is the winner.
Take a quick peek at Wikipedia and you'll see that the technology has been used with some success in a variety of situations, such as predicting the outcome of presidential elections, Oscar nominations, and sporting events.
Could the predictive market approach be useful to storage? John Ives, president and co-founder of Storage Markets , and Vice President and co-founder Rich Pappas have no doubts.
"This is really a tough industry," Pappas points out. "If I am a semiconductor company, today I am making funding decisions that [will] impact my product line up to three or four years from now."
To help support those decisions, Ives and Pappas founded Storage Markets and collected a group of traders with diversified expertise in storage. Here is a chart of the current traders demographics by type of company, but that might change as more traders join.
Vendors are obviously the primary customers of Storage Markets' services, and will provide questions that traders will answer with their transactions. Here is an example of how a question about two disk drives models reaching the same price could be answered.
As an honorary trader, I can testify that Storage Markets is fun and addictive, which should motivate more people with storage expertise to join and keep the transactions going. The company is not discussing prices, but their services should be competitive with alternative market research options, according to Pappas.
End users may have something to gain from all this, too: If vendors save money by basing their decisions on more accurate market predictions, that should translate into less expensive goods.
Am I being too optimistic? If you have storage expertise, try becoming a trader at Storage Markets and start betting on the future of storage.
Posted by Mario Apicella on October 30, 2006 05:40 AM
September 28, 2006 | Comments: (0)
Well you should, by now, at least according to this article by Seagate's Willis Whittington.
Sorry for the ricochet but the Whittington piece is very good, perhaps should be mandatory, reading for anyone interested to learn more about the successor of parallel SCSI.
As alway the comments box is right below. Let me hear from you and thank you for reading.
Posted by Mario Apicella on September 28, 2006 12:06 PM
September 14, 2006 | Comments: (0)
50 years of hard drives: It's time to move on.
Maybe it's because I am not much for celebrating anniversaries, especially those related to unanimated objects, but isn't all the hubbub around the disk drive hitting the big five-o getting annoying?
The disk drive was a clever invention, I won't deny that, but so were others. Did anyone ever throw a party for, say, the fiftieth of the printer? Well, seems to me that printers predate disk drives in the computing ecosystem but got no party.
What about monitors, or keyboards or modems? Aren't those machines at least equally important, not to mention more reliable, than disk drives, those irritating recycling bins for bits of data?
Sure, computer memories based on disk drives (we call that storage now) provide open-ended capacity and quick access, but what about persistence? Among all media the disk drive is one of the least reliable for long term storage, unless you do recurring re-fresh cycles on newer drives, but that's like cheating, isn't it?
We know that paper can survive centuries, in fact our museums have many examples of precious antique books still in perfect shape. However we cannot entrust important data to our spinning friends, not even for a few seconds without taking special precautions.
So, why do we celebrate such a poor performer? Is this becoming a way of life these days? Celebrating failure, I mean.
Perhaps because, if I stop pretending to be naive for a second, an 800 pound gorilla like IBM can claim paternity of this much discussed child? Well take a look at that first baby. Not so pretty, don't you agree?
In fact, the only reason I am publishing this link is to prove a birth date, September 4, 1956 according to ma' IBM, which is coincidentally also my birth day, but not on the same year, mind you.
Could another reason for this disk celebration craze be that although Big Blue is no longer deeply involved in filling planet Earth with disk drives, their prestigious research labs are still trying to make a better trap for bits of data?
The answer to both questions is obviously yes. In fact, as part of the celebrations last week IBM invited reporters to a day long event titled: "The future of storage" at the Almaden Research Center in California.
The event's main course was an interesting peek at current research topics. It's difficult not to get excited at the promise coming from those technologies.
Take, for example, SCM (Storage-Class Memory) a solid-state alternative to disk drives (and flash memory) with faster access, better reliability and less draining power and cooling demands than today's spinning devices. Look ma', no moving parts!
Gian-Luca Bona, department group manager of Science & Technology at Almaden explains how the cost of any memory medium is directly affected by the size of the bit cell, where larger equals more expensive.
Trying to compete with the density of magnetic media is just one of the many challenges that researchers are facing. According to Bona, Storage Class Memory could become ready for mass production in just a few years, but he is quick to add that other, non-technical factors come into play when a possibly disruptive technology like this pops up from the labs.
I'll keep a bottle of champagne ready for when the disk drive will finally find its match. Considering that many of the storage problems that we face today are more or less directly related to the characteristics of the disk drives, I for one will celebrate when we will finally turn page.
That celebration won't happen soon, I am afraid, and we will continue to pile up disk drives and their problems for the foreseeable future.
Meanwhile, I hear from Seagate that the storage areal density has just reached 421 gigabits per square inch, a new record. I used to get excited at these announcements but I can't now: Aren't we just feeding our addiction to larger and more troublesome disk drives?
Posted by Mario Apicella on September 14, 2006 05:48 AM
July 31, 2006 | Comments: (0)
Giving Windows a boot to iSCSI
Last week emBoot quietly released the final winBoot/i, a product that was first announced last Spring at SNW. A copy of the release is here.
The product was delayed waiting for Microsoft to release the required iSCSI software initiator versiontells me Mickey McIntire, former CEO of String Bean Software (remember Wintarget?) a company that was recently acquired by Microsoft.
What has McIntire to do with emBoot?
With more time on my hands I've started getting back into the consulting business with my first client being emBoot, Inc.he explains.
I couldn't resist downloading an eval copy of winBoot/i and you should do the same if you have iSCSI storage in a Windows environment.
How does it work? You install the winBoot/i server on one of your machines, typically where your DHCP server is. Then allocate some boot image volumes from your iSCSI target, usually one for each Windows server you want to protect.
Moving to the candidate servers ,install the client version of winBoot/i which will install also the boot capable version of the Microsoft iSCSI Initiator. As last touch, using the iSCSI initiator at each server grab the correspondent boot image volume created before.
Now you're ready to make a copy of the boot disk of each server to its iSCSI volume. Appropriately, emBoot gives you a choice to use the installed winBoot/i client or to boot from the installation CD and save the pain of managing those open files.
This is what you should see when booting from the CD to make a copy View image.
That's it, set your servers to boot from the NIC and you should be ready for diskless servers. I am still testing, so don't have reached a final opinion yet, but I like what I've seen so far.
However, it helps to have at least GbE on each machine even better 10GbE, if you can squeeze that in your budget.
We have found the vast majority of GbE NICs in use are Broadcom or Intel. Broadcom NICs, we have found on many occasions, need the very latest in PXE and NDIS code to avoid freezes on startup to Windows iSCSI boot. We've injected their very latest drivers into the Windows PE-based Client Deployment Tool [bootable CD or PXE image].tells me Steve Marfisi, Manager of Technical Services for emBoot.
Of course, every environment has its own potential bumps in the road, which is why the free trial offered by emBoot makes much sense.
We've also added support for Neterion Xframe 10 GbE NICs and Alacritech iSCSI NICs, as well as VMware NIC driversMarfisi explains.
The client license for winBoot/i is $95 for each machine. The license for the first winBoot/i server is free, but you'll pay $395 each if you need additional servers.
It's tempting to say that you could cover the cost of deploying winBoot/i with the cost of the drives spared on each server. However, I won't say that because the main reason to boot your servers from iSCSI volumes should not be to save money on drives, but to give your data center a more manageable and forward looking structure.
Isn't it about time for your OS volumes to become part of a SAN? I think it is and winBoot/i gives you the tools to do exactly that.
Very high on my "check this out" meter.
Posted by Mario Apicella on July 31, 2006 07:52 AM
July 27, 2006 | Comments: (0)
I want to share the comment of a reader to my Storage Insider column on Sony announcing a Blu-ray recorder and possibly have also your comments on that.
Not sure what I am talking about? Please check this out and speak out your mind.
This is what Bill Davidsen had to say (forwarded to me via email):
The main attraction of Blue-Ray is that the media is not light sensitive, not magnetic, reasonable cost vs. tape solutions in the 200GB-2TB range, and available before the end of the year.HD-DVD seems aimed at the entertainment market, Blue-Ray seems to be a better fit for archival storage.
Do you agree with Bill? Not at all? Use that comment window below...
Posted by Mario Apicella on July 27, 2006 12:38 PM
July 07, 2006 | Comments: (0)
Mirror, mirror on the wall, who makes the fastest file server in the land?
If someone at NetApp was actually echoing Snowhite's wicked stepmother while checking out the benchmark results posted at spec.org, that person was probably not liking the answer.
In fact, until now systems from other vendors including BlueArc, EMC, Exanet, HP, IBM and Panasas had produced more operation per second than NetApp's running the aging SPECsfs97 benchmark.
With the recently announced Data Ontap GX, an offspring of the Spinnaker acquisition, NetApp has finally put a system at the top of the list.
Just for fun I put in a table the top ten results, listing the number of operations per second, the overall response time and how many drives and how much memory were on the test systems.
Enjoy, but please keep in mind that the SPECsfs97 benchmarks produce much more data than I could list here. For example, knowing how many file systems were used, the characteristics of the network and the number and type of CPUs could make a big difference in your evaluation.
Have fun as I did, comparing results and settings.
| Ten fastest NFS file serving systems according to SPECsfs97 published results as of July 5, 2006 | ||||||
|---|---|---|---|---|---|---|
| Company | System | Oper/sec | Overall Resp. time | Disks | Memory | Date published |
| Network Appliance, Inc. | Data ONTAP GX System (24-node FAS6070) | 1,032,461 | 1.53 | 2016 | 768 GB (32 GB per node) | Jun-06 |
| EMC Corp. | Celerra NSX Cluster 8 X-Blade 60 (1 stdby) 2 DMX | 320,064 | 1.64 | 770 | 28 GB (4 GB per blade) | May-05 |
| Panasas, Inc. | Panasas ActiveScale storage cluster (60 DirectorBlades) | 305,805 | 1.76 | 540 | 240 GB (4 GB per DirectorBlade) | Dec-03 |
| Exanet Inc. | ExaStore EX600FC | 203,182 | 1.08 | 348 | 72 GB (12 GB per node) | Oct-04 |
| BlueArc Corporation | BlueArc Titan 2200, 2-Node Active/Active Cluster | 195,502 | 2.37 | 416 | 47.6 GB (not including other_cache_size and NVRAM) | Feb-06 |
| Exanet Inc. | ExaStore EX600FC | 178,156 | 1.04 | 348 | 72 GB (12 GB per node) | Oct-04 |
| IBM Corporation | IBM System p5 570 (2.2GHz, 8 CPU, TCP) | 169,786 | 1.67 | 421 | 128 GB | Feb-06 |
| IBM Corporation | IBM eServer pSeries 690 (1.7GHz, TCP) | 167,007 | 1.08 | 561 | 128 GB | Oct-03 |
| Hewlett Packard | HP AlphaServer GS1280 Model 16 7/1150 | 154,654 | 1.15 | 337 | 96 GB | Apr-03 |
| IBM Corporation | IBM eServer p5 570 (1.9GHz, TCP) | 145,362 | 1.17 | 561 | 128 GB | Aug-04 |
Posted by Mario Apicella on July 7, 2006 08:34 AM
June 02, 2006 | Comments: (0)
WP from Infortrend: why SATA plus RAID5 puts your data at risk
I have reached the conclusion that I'll never see the end of my reading backlog: There are so many interesting topics and unfortunately so little time. If anyone has found a way to read (and understand) things faster I am all ears.
Regardless, part of my catching up with the world this week was reading a white paper on RAID6 from Infortrend that I wholeheartedly suggest reading to anyone is using or thinking to use SATA drives.
Why? Because there are some risks involved in using RAID 5 with those large drives that could affect the integrity of your data.
Wait! I was going down too lightly with this, let me rephrase: YOU COULD LOSE ALL YOUR DATA using RAID5 and SATA drives.
You may have heard this before, (and so did I) but still should read that paper. It's not exactly a new topic, but one certainly worth understanding better.
Which brings me back to the WP from Infortrend that explains the why and the how with detached accuracy and professionalism, and using the most effective human language, math. For those who are not comfortable with equations, the WP has crystal clear tables and charts.
Update: The author of the WP is Ted Pang, director of technical research for Infortrend.
Once again, RAID5 plus SATA is a bad idea. Actually, I will go out on a limb and say RAID5 plus ANY large drive is a bad idea. Do yourself a favor and read that WP from Infortrend to find out why.
Now, what's my next reading assignment?
Posted by Mario Apicella on June 2, 2006 06:38 AM
June 01, 2006 | Comments: (0)
LSI Logic ships low-profile SAS controllers
LSI Logic ships low-profile SAS controllers
Posted by Mario Apicella on June 1, 2006 12:14 PM
March 06, 2006 | Comments: (0)
Microsoft gets on target with iSCSI
At the end of a rather busy week for storage (is there any other kind, I wonder?), Microsoft announced the intention to acquire WinTarget from String Bean Software and to make that technology part of Windows Storage Server in the future, starting with WSS 2003 R2.
What will Microsoft gain from adding WinTarget technology to WSS? In essence, the ability to serve both blocks and files to its clients over an iSCSI network.
I tried WinTarget long time ago, and more recently Neterion and String Bean Software published some interesting benchmark results measured on machines equipped with 10Gb Ethernet adapters and WinTarget.
Back to Microsoft, adding iSCSI target capabilities to WSS will open some interesting opportunities for server vendors that will be able to add more spice and more value, not to mention a few revenue dollars, to their machines.
WinTarget makes a perfect complement to other Microsoft technologies including of course the iSCSI initiator and the NAS capabilities of WSS.
I would be surprised if server vendors would not jump on that opportunity to differentiate their products from competitors' and to lure customers with the promise of a fast and affordable iSCSI SAN.
To quote Mickey McIntire, CEO of String Bean Software and esteemed contributor to this blog:
iSCSI SANs have crawled a long way up the evolutionary ladder in a short period of time and are now encroaching on the previously unchallenged domain of Fibre Channel.
Posted by Mario Apicella on March 6, 2006 07:59 AM
August 26, 2005 | Comments: (0)
Desktop class RAID has been a PAIN for many years
Says Michael Lueck from Lueck Data Systems in response to this Storage Insider column: Simplifying SATA RAID
But let's hear from his own words.
Greetings Mario-
It will be interesting to see how Silicon Image plays out this "new" desktop RAID chip technology.
Desktop class RAID has been a PAIN for many years now. PCI-X is not helping matters any as desktop class boards have 1x and 16x slots, with the 16x taken by the graphics boards. Real SATA RAID boards need an 8x slot which is only found on dual processor workstation / server class boards... and then the graphics typically stink.
I say it will be interesting to see how they deal with the RAID for desktop class as there is quite a bit of RAID out there on desktop boards, but it is always this BIOS tricking Windows-centric junk that I would not trust anything to, and like I inferred already is Windows-centric.
To be fair to Silicon Image, their chip is OS agnostic but the OEMs could decide that Windows users are a more attractive target. In my opinion, setting up a RAID card should be a no brainer for a person smart enough to install and manage a Linux desktop.
But let's put Michael on again.
I end up having to stick a 64-bit 3ware controller in a 32-bit slot for now to get a low cost SATA mirror into low-cost firewall boxes.
Being a performance nut, I hate seeing 32-bits hang into thin air, but vendors have really backed the industry into a corner of restricting real hardware RAID away from desktop class computers.
I am as anxious as you are to see this new chip at work. Mirroring (RAID 1) should become a minimum requirement on all laptops and desktops. A single disk drive should not be trusted to reliably store any single-copy file.
Thank you for those provocative comments, Michael.
Posted by Mario Apicella on August 26, 2005 07:55 AM
June 20, 2005 | Comments: (0)
When I first read about this on Physorg.com I was interested but rather skeptical.
Let me clarify that Physorg is one of my favorite and most appreciated readings, and that my skepticism was mostly because of the topic and not because of the source.
After all, you don't read every day that scientists are developing a way to recreate matter over an Internet connection.
Shocked? Well I was too. Let me repeat: in a not too distant future, we should be able to capture the essential characteristics of a local object, translate that into a data profile and send it to a remote place.
Nothing out of the ordinary, so far, but here's the scoop: reversing the process at the remote site we would create a credible replica of the original from the data profile. SciFi anyone?
Well, it's not SciFi and it's not woodoo magic either, but an actual project driven by Seth Goldstein & Todd Mowry, two eminent professors at the Carnegie Mellon School of Computer Science.
The main ingredient to achieve that phenomenal alchemy from data to an actual object is a new element that the two professors call programmable matter.
Care to read a definition of programmable matter? here it is, from the horses' mouths:
An ensemble of material that contains sufficient:
– local computation
– actuation
– storage
– energy
– sensing & communication
Which can be programmed to form interesting dynamic shapes and configurations.
It all starts to make sense now, does it not? Now close your eyes and visualize these atoms of programmable matter automatically joining to form objects consistent with the original profile.
The process as I understand it should be very much similar to what happens in a chemistry lab experiment, with the difference that atoms of programmable matter will follow their own programming rather than the laws of conventional science.
Programmable matter could become reality in five years or less according to the project, but with rather large atoms, which would make for a rather coarse rendering of the original object. In time those atoms should become more finely grained.
Granted, you won't see a review of products based on programmable matter anytime soon, but the perspective of this research producing a prototype in 10 years or less gets me going.
Just imagining possible applications of this technology is mind numbing. Think for example of sales reps delivering their pitch while creating accurate replicas of their goods.
Or online shoppers, downloading the profile and creating a proxy object before committing to a purchase. How about trasmitting the profile of a real human being? Can you imagine being able to "download" your favorite actor? Gives an all new meaning to the expression "home theater", does it not?
Before getting lost in reveries let me suggest a more sobering thought: if we have so much trouble managing static, two dimensional data, how will we adapt to the mercurial characteristics of programmable matter?
Posted by Mario Apicella on June 20, 2005 11:48 AM
April 20, 2005 | Comments: (0)
Moving toward object storage devices
If you walked by the Seagate booth at the recent Storage Networking World in Phoenix, you probably noticed as I did an interesting demo of object storage technology.
To a casual observer the demo was nothing too glamorous: a couple of IBM servers sending data to storage devices over Ethernet and, via an Emulex HBA, to an array of Seagate drives.
Well, to a casual observer Neil Armstrong's historical landing on the Moon was probably just a step like many others, but we know better.
Indeed, this not much rumored public display of object storage technology could mark the beginning of a significant leap forward for this industry.
At the booth I had the pleasure of meeting Erik Riedel from Seagate Research, who gave me a captivating update on object storage.
Riedel chairs with IBM's Julian Satran the Object-Based Storage Devices Technical Workgroup at SNIA, and has a privileged view of the state of this technology.
The concept behind object storage is rather simple: to give storage devices more control over basic features such as allocating sectors for a file. Obviously, in an object world the OS does not manage those activities anymore, which translates into increased transparency of a device data content across different OSes.
Although the smaller unit of data for disk drives is still (obviously) the sector, the minimum exchange unit between OS and object devices becomes, you guess it, an object.
For example, to store this article the operating system would pass the whole content as an object to the storage device, receiving in return an object ID to associate with its filename.
To retrieve the same article, the OS would find its filename in a directory, pass the associated ID along with a request to retrieve an object to the device, and grab the results.
If you are thinking that to accommodate object storage devices the set of SCSI commands needs to be expanded, you're absolutely right. But the impact of object storage could go much further.
Think of how many other activities besides space allocation an intelligent object device could manage independently from the OS. Consider for example, access control, encryption, replication, criteria-based data movement across devices .. the possibilities are nearly endless.
Perhaps the most immediate impact of object storage will be to finally settle the dispute between serving block and serving files. Of course, the winner will be serving files, or to be more accurate, serving objects.
Other possible implementations of the technology will probably be decided, as usual, by criteria of efficiency and cost. Regardless, object storage promises to be the greatest revolution since networked storage.
Posted by Mario Apicella on April 20, 2005 10:01 AM
March 17, 2005 | Comments: (0)
It's great to see a technolgy like SAS (that's for serial attached SCSI, not the other stuff) mentioned almost two years ago in one of our first columns on the topic finally come to market.
In that column the actual availability date of SAS-based products was off by a long shot, but other promises of the technology seem to be right on target.
It may be worth reading this news article from IDG in its entirety, but, to give a quick summary, HP is getting ready to offer SAS drives in its products, possibly as soon as May.
While this is a worthy scoop in itself, perhaps the most interesting part of the story is that HP will offer SFF (small form factor), 2.5" SAS drives in its servers, blade systems and storage arrays.
HP is the first to make a public announcements about SAS products, but it's reasonable to assume that other vendors will follow suit because this technology open opportunities that are too good to pass on.
At this early stage neither HP nor other vendors are making public product specs updated for SAS, but we can still do some guessing on what to expect.
Some of the benefits of mounting SFF drives are easy to guess: less noise, less heat produced, hence enclosures that are easier to cool, and obviously more capacity per rack unit.
Now let's weave over that picture another layer of major benefits, these typical of SAS , such as smarter connectivity on each port, with either a single point- to-point connection or a fan-out to multiple devices. This flexibility of SAS ports can be exploited to build enclosures with high capacity or fast data transfers.
SAS has small cables and small ports, which makes possible building two connections on each 2.5" drive. Link that dual path to two different controllers and you attain redundancy that has nothing to envy to high end FC (fibre channel) solutions.
Not less important, SAS and SATA (serial ATA) drives can attach to the same controller, which gives customers the option to size the best mix of capacity and performance in each enclosure.
SATA drives are less expensive and probably will remain more capacious than SAS, which means that an array with mixed drives will be less expensive than one with just serial SCSI.
So, judging from these benefits, what should we expect from the news SAS products? Smaller enclosures? Larger density? Less environmental burden? More reliability? Less expensive storage?
Probably the correct answer is "all of the above" but customers, as usual, will have the ultimate say.
Still hungry for SAS? Try this article written by Tonya Comer from HP for the SCSI Trade Association.
Posted by Mario Apicella on March 17, 2005 08:45 AM
February 10, 2005 | Comments: (0)
Cost, complexity the main obstacles to blending medicine and IT
Says Brad Eacker, in response to the IT's next challenge: healthcare. Brad has quite a record, both as consumer of healthcare services and as IT guru. But let's hear from him.
Mario,
Having dealt with both healthcare (as a patient) and storage (as a developer) for many years, your articles and blog entries regarding this crossover resonated with me.
On the storage side, I was responsible for putting together a 200 GB Oracle database in '94 as a proof of concept for SGI, utilizing 250x2G drives. My job duties also had me developing an I/O system for the SGI Challenge architecture which provided over 510 MB/sec disk to memory throughput bask in '93. Some of my current work involves a 900GB Oracle database conversion to a MySQL environment. So large data sets are part and parcel to my work experience (24 years in software development).
On the health side my body has been through a Harrington rod implantation procedure for scoliosis repair ('76), an arthroscopic procedure to reattach an CAL that tore off a chunk of my tibia ('90) and a craniotomy to attend to a grade III oligoastrocytoma ('99). These procedures have resulted in my body having experienced a multitude of medical tests, surgeries, x-rays, MRIs, blood work, pulmonary tests, radiation treatment, chemo-therapy, and even halo-femur traction for 2 weeks.
So I can clearly understand both sides of this situation. In that understanding some elements have been standardized, mainly for the medical insurance needs, ICD-9 is the current coding mechanisms from the CDC with ICD-10 coming slowly into play.
So there are some attempts by the US Government to attend to the standardization you've alluded to in your article. But not nearly close enough to what will be necessary to fully support the HIPAA portability desires. And trying to get the doctors to actually utilize even some of the possible online echanisms will require an ease of use that just doesn't exist at the moment.
I can envision an XML type, self describing data format as one possible mechanism for how to deal with the vast array of classifications for doctors, medical procedures, opinions, observations, prescriptions, and all the other elements that attempt to describe an individual's medical history in a portable fashion. But such a mechanism hasn't been a real target for the healthcare environs.
Their bottom line doesn't really support the necessary development of the systems to facilitate this type of communication and portability. And any kind of direction coming from the federal government tends to take on the classic 'unfunded mandate' moniker.
Hopefully the various parties will start to understand that the information gathered during the course of an individual's life regarding his health situation can lead to better care, less mistakes, and a better understanding of how a body reacts to all the machinations that the body and mind go through during the course of a life. Such knowledge will allow others to learn from the lessons previous patients have already had to deal with.
Brad Eacker
DBA, kernel developer, technologist, and patient
Thank you for sharing those thoughts and your experience, Brad. Stay healthy.
Posted by Mario Apicella on February 10, 2005 07:50 AM
| STORAGE SPRAWL PODCAST |
| Listen to the latest podcast: |
| MP3 • RSS • Get Podcast • Archive • Mobilize |
TOP STORIES
ADDITIONAL RESOURCES
- Network Storage: Patterns, Trends & Business Needs
- Application Security: Threats and How to Counter Them
- Remote Access: Maintain Security and Decrease the Burden on IT
- Disaster Recovery in Minutes
- Reduce Recovery Times and Tape Costs
- Secure, recover, and archive critical information
