Save $15 On Any Dreamhost Plan: coupon "stomojo"
From the Networker 7.3.2 Service Pack announcement:
• Support for Sun ZFS for improved support for customers running Solaris 10
We’re not actually endorsing ZFS you see
This is just to take care of our many loyal Solaris 10 customers.
I don’t care why they do it - it’s all good.
Zmanda: start your engines!
Thanks to James Dickens for spotting this.
Several companies are in the business of providing software that turns a USB thumb drive, or iPod, into a simple way to carry your own data and programs for use on any PC. Today, one of them Ringcube, hit the PR jackpot with a largely favorable review in the Wall Street Journal’s Mossberg Solution (paid subscription required) column.
A Brilliant Name, A Good Product
Ringcube’s product, MojoPac, is software that enables you to load a drive with your data and your favorite programs and take them on the road. As Mossberg put it:
MojoPac sounded too good to be true, but for the most part, it actually worked as promised. Privacy is a big plus for MojoPac, as your files remain on your thumb drive or iPod, and never transfer to the host PC’s hard disk. Similarly, your entire browsing history and all cookies remain on the portable device.
There is a catch, however: A few aspects of this program are a little too geeky for the average person, it is slow to perform some tasks, and it crashed one of our computers during a test. Also, it doesn’t support making Microsoft Office portable, unless you have a corporate or institutional license.
But the company claims it is hoping to make the geekier parts of MojoPac more user-friendly in its next software update, and is working on allowing average consumers to carry their copies of Office with them. MojoPac only works with Windows XP programs as of now.
What about Ceedo, U3 and Migo?
I noted Ceedo a few months ago, and looked at but didn’t write about U3 as well. Both of these have similar functionality to MojoPac with one major exception: they require use of specially adapted programs or a special, extra-cost software adapter. Many of these programs are free, but many are not, and that would be a deal breaker for me. U3 also requires a U3-logo’d USB drive. And Migo doesn’t run programs, just stores data.
So far, MojoPac seems to be the best PC on a stick product out there. It allows you use the applications you are used to, costs less than U3 and Ceedo, runs on any USB storage device and, according to Mossberg, seems to have the easiest UI for many of its functions. On the downside, Mossberg did report some bugs, so it probably isn’t as well wrung-out as the older solutions.
Comments welcome, as always. Moderation is turned on to deter comment spam and registration is turned off to simplify commenting.
Gear6 goes semi-public today with their new network storage cache, saying what the product does, but not releasing configuration, pricing and other key details until later. So some of my questions, and probably yours, remain unanswered. Here’s the Cliff Notes version.
Network storage cache cluster
It is a network cache appliance for, initially, NFS filers. Plug it into the same ethernet switch the filers are on, perform some voodoo at the Gear6 management interface to assign filers to be cached, and you now have a really large - how much joy can you afford? - and scalable cache sitting between the servers and the filers.
This isn’t a RAM disk, because you don’t need another disk to manage. Gear6 says they have significant IP in the tools that automate the populating of the cache.
Why bother?
Gear6 says their beta customers work across many industries and applications with a few common characteristics:
If this sounds like you, check out Gear6.
Adding capability to the network
Three months ago I did a once-over-lightly competitive analysis of Gear6 and concluded they were building a
. . . honking-fast Linux-based parallel MIMD FC-SAN non-blocking I/O NAS appliance designed to handle hundreds of thousands of concurrent I/O’s from large numbers of servers, clustered or not. . . . highly scalable, you’ll be able to add processors and network interconnects just as you would in any cluster or grid. [No] custom hardware, preferring to use very smart software, such as the CxC parallel computing language, on commodity clustered servers and interconnects.
So it’s Gig E instead of FC, and millions of IOPS, but overall I ‘d give myself a B+ or even an A- for that prediction.
The StorageMojo.com take
Blowing functionality across the network is a long-term secular trend. By focusing on building an NFS SAN cache appliance I believe Gear6 has picked a niche ripe for change. Depending on their pricing and management mojo it may make sense for F1000 customers to start buying really cheap filers with minimal cache and let the Gear6 cache cluster do the heavy lifting.
Comments welcome, of course. Moderation is turned on to control comment spam, but no registration required.
In part I some of the reasons the current model of enterprise IT is broken were explored. In the concluding part, I make a radical suggestion about how IT can be fixed - and why Internet Data Centers are better positioned to implement the fix and perhaps force enterprise IT into a long slow spiral of decline.
This section discusses why Service Oriented Architectures (SOA) are a long-term secular trend:
Accelerating demand for IT services
Rapidly growing IT resource demand show no signs of slowing and may even accelerate due to the advent of automated web services. The growth results from the mass commoditization of supply and the automation of demand: vastly increased supply at much lower prices combined with increasing economic demand for easily accessible information. Web 2.0 - or whatever you want to call it - automates IT demand well beyond anything we’ve ever seen.
The fundamental economic driver of increased IT demand is that electronic data is cheaper and easier to access, process and store. The ease and low cost enable higher quality services and better decision making, pushing competitors to exploit the technology in order to keep up. The rise of E-commerce has dramatically increased the volume of easily collected business data. Web services will further increase the volume and scope of such data, as well as further pushing business into 7/24 operations since machines never sleep. In addition the continuing decrease in infrastructure and programming costs makes it economic to automate business services that were too costly even five years ago such as
Commoditization of IT infrastructure
Commoditization of key infrastructure is driving and will continue to drive IT technology costs lower than even Moore’s law would suggest. As IT buyers concentrate their spending on fewer architectures (CPU, network, storage, database, OS etc.) these technologies get the benefit of increased volume and lower cost if there is competition. Hardware costs have benefited most, but even software pricing has been affected.
The widespread adoption of Linux has hastened the commoditization of operating systems and brought the industry down to two basic camps: Unix and Windows. Pre-packaged software and sophisticated development tools have commoditized most application and development software. Even twenty years ago many companies wrote their own business applications for inventory, order entry and other critical business processes. Today the focus is on integrating across applications using tools and standards, such as XML, that were unimaginable 10 years ago.
The commoditization of the IT infrastructure means it is feasible for the first time to build SOA infrastructures. By reducing the number of variables (storage, server, network, OS and applications) infrastructures are now ripe for the kind of centralized provisioning and management that telcos have had for decades.
Automation of all business processes (Resiliant infrastructure now required)
The virtuous cycle of lower cost infrastructure enabling the automation of more business processes thereby driving demand for IT products continues unabated. As more business processes are automated, the infrastructure is expected to handle greater variation in demand at the same time that uptime becomes more critical. Most systems even 10 years ago had manual backup processes that could handle IT outages of a few hours to a few days. Today however few companies have enough experienced employees to buffer IT outages. The automated systems have to work or business simply stops until the system gets fixed.
Flat budgets and exploding IT demands will push more enterprise executives to consider an SOA model – whether services are deployed internally or entrusted to an outsourcer. Not all enterprise data is equally sensitive so the LOB’s will begin testing the waters with less critical applications that debug the infrastructure. Internal IT organizations are simply another service provider. Expect LOBs to invite external service providers to take part of the business to provide a competitive benchmark. The growing economic pressure to move to the SOA model redefines the infrastructure challenges whether IT services are provided by a captive internal organization or an external supplier.
Resource Allocation in Service Architectures
Service architectures change the way resources are allocated – moving from the element to the event. Events define service: a server added to a cluster; a logical volume capacity increase; a packet sent or received; a cache hit; or a database access. In fact any event that can be captured could be used to help allocate resources. Tie these events to specific resources and users, through source and destination addresses, system or file names, and the event is placed in its economic context: who generated the event and what was its impact on the infrastructure. This algorithm contrasts sharply with the taxation model that supports enterprise IT today.
The SOA pricing model
Traditionally a LOB purchases the elements required to run a given application (server, storage, database and application software) and IT would operate the system. In service architectures the elements change rapidly: an application might require 6 blades and 100 GB at the end of the quarter and only 2 blades and 50GB a week later. Now the LOB is purchasing the service provided by the infrastructure, not the infrastructure itself. If the resources are going to vary based upon demand the best way to measure and pay for them is through events rather than ownership.
In practice some elements are more fungible (interchangeable) than others. Network bandwidth is fungible: a packet is a packet, and availability is similar across the enterprise. Servers and storage are less so, especially at the higher end of the market. While 1U servers are reasonably similar, an MVS mainframe is quite different from a Solaris Sunfire server, even though both may be running Unix. Rather than attempt to ignore the differences a rational resource allocation system would reflect these differences so the LOBs would have the incentive to maximize the business benefit of the various systems.
The SOA advantage
In The Capacity Illusion one of my commenters noted that the intelligent thing to do is to waste what is cheap in order to preserve what is expensive. Tax-based funding - what enterprise IT has today - fails in two dimensions:
Taxes vs. a free market
The most obvious alternative to paying for corporate IT with taxes on the LOBs, is a free market, where the LOBs order services from either internal or external IT providers. This is where IDC’s such as Amazon have a key advantage: they already charge for their services in a fairly granular fashion, and are highly competitive to boot.
Markets aren’t perfect for allocating resources, but over the long-haul they’ve proven to be the best mechanism we have. The big advantage of SOA’s that they are well suited for event-based pricing and enterprise IT isn’t. This goes well beyond today’s clumsy chargebacks that peanut-butter costs across a few areas, such as storage or network accesss. Pricing needs to be as granular as possible so that high-value services cost more and low-value services cost less.
Enterprise IT, if it saw the need, could turn itself into a price-based SOA. Current cell-phone rating software, which routinely handles millions of events with thousands of pricing models could be adapted to handle enterprise SOA pricing. The technology exists to turn enterprise IT into a bazaar instead of a cathedral - to steal a phrase from Eric Raymond. A bazaar where LOBs finally have a reason to take responsibility for their IT usage and where IT gets unambigous feedback on what is important, measured in dollars.
Does there need to be a part III? Comments welcome. Moderation is turned on to control comment spam, and registration is turned off so you can comment faster.
Putting out my shingle
I’ve been consulting for several years now. I love it because I get to help people solve a lot more interesting problems than any single job I’ve ever had.
Since I’ve chosen to live out in the boonies (see photo above), I launched StorageMojo.com to help me stay in touch with my favorite technologies. The discipline of researching and writing about data storage, in all its guises, five days a week has been more rewarding than I’d dared hope.
You’ve read the blog . . .
If you like how I frame problems - or how my independent perspective inspires fresh thinking - you already understand much of what I bring to the table. Click on the Consulting tab above and learn more.
Fighting rear-guard marketing actions - protecting a declining product or technology to keep the high-margin revenue coming in - isn’t much fun, but it sure is profitable. Especially in an area as conservative as storage. So it is a big deal when someone quits trying.
Which is why I’m so [shocked/surprised/pleased] to see Gudmundur Einarsson, CEO of Tandberg Data face reality and admit the facts in a press release:
“Using RDX QuikStor, organisations that currently rely on tape to protect, archive and interchange their data, will benefit from enhanced performance, instant random access, high reliability and, most importantly for small businesses, a lower cost than competing low-end tape solutions. . . . ”
Low-end RDX product - but can the high-end be far behind?
The RDX product packages cheap, 4200 RPM, 2.5″ disks in what T’berg calls a rugged, shockproof cartridge with a 10 year life in 40, 80 and 120 GB sizes. The unit starts at about $300, so the margins look good for T’berg and the resellers should be able to move lots to the SOHO and SMB markets.
Calling high-end tape vendors
Tandberg stresses that the RDX is a low-end system. Yet people have been predicting for years that disk would overtake tape. How close are we? As the table shows, pretty darn close.
| SuperDLT IV Tape | 300 | ~$85 | $0.28 |
| SATA Disk | 300 | ~$100 | $0.33 |
| LTO Ultrium 3 Tape | 400 | ~$60 | $0.15 |
| SATA Disk | 400 | ~$160 | $0.40 |
Alert readers will notice that DLT is much costlier than LTO, reflecting the premium tape cost that includes Quantum’s lucrative licensing fee and the inability of an aging architecture to keep up with the newer LTO. Which is why DLT is well on its way to that big archive in the sky.
I’ve also left some signifcant costs out. To put an HDD in a removable enclosure runs about $40, quantity 1. And I’ve left out the cost of the tape drive, which for these top of the line Ultrium 3 starts north of $2,000. Also, I don’t take into account the 2:1 compression manufacturers commonly assume, and that customers may or may not see.
The StorageMojo.com take
At the rate disk prices drop, this picture will look very different in 6 months, when the 1 TB drives are shipping and the industry shift to perpendicular recording is complete. And if the folks trading at Storage Markets are correct that we will see GB price parity next year, then 2.5″ will become very attractive as a backup medium when properly packaged and priced. The tape business is beginning to unwind, and while it won’t go away any time soon at the high-end, it brings almost nothing to the SOHO and SMB market space.
Expect to see the D2D data reduction companies do very well over the next three years.
Comments, as always, welcome. Moderation is turned on to keep out comment spam, and registration has been turned off to make commenting faster and easier.
Why do CIO’s have the shortest average job tenures of any C-level executive? The simple answer is because users (Line-of-Business (LOB) managers) aren’t happy with them or the performance of their organizations. Many issues contribute to this situation, but the bottom line is that CIO’s are not meeting the expectations of their peers and they get whacked for it.
CIOs offer some reasons for this dissatisfaction such as:
and other variations on these themes. These are all issues but they are also excuses.
Other C-level executives deal with difficult processes and environments they don’t control, unrealistic demands and resource constraints. Yet somehow they persevere and largely succeed. LOB frustrations with IT are seen in periodic efforts to bring IT into their business units to escape IT control of vital business services. Inevitably however, the LOBs want IT to take over the IT function again as they confront the issues of scale and expertise required to manage modern IT infrastructures. From an LOB perspective the relationship with IT is like the old saw about relationships between the sexes: “you can’t live with ‘em and you can’t live without ‘em”.
The IT crisis spans all management styles, organizations, industries, funding methods and technologies. The issues between LOBs and IT are not of personalities or organizational design. The basic model of IT is broken. Specifically, there are three key problems:
Measuring enterprise IT value-add
It is a paradox: modern enterprises run on IT systems; but the value of those systems is almost impossible to quantify. As a result, IT cannot business justify their capital budgets against the business results those systems generate. Certainly, new applications sponsored by the LOBs get supported, but how does the enterprise know if it is spending too much, too little, or just the right amount on IT? It doesn’t. And IT doesn’t know how to get the answers that executives need.
IT is a service that supports the revenue-generating businesses of the enterprise. As a service, IT can’t tell the business units what IT services are worth. Only the LOB’s can do that. The LOB knows its customers, its costs and its margins, so the LOB is the logical entity to assign value to IT services. Yet due to the way IT services are funded the LOB has no more idea than IT what IT services are worth.
IT commoditization and Service Oriented Architectures (SOA)
SOA is the latest big organizing principle for IT architecture and product development. In the past it was mainframes, distributed computing, client-server and web-based in the 1960’s, 70’s, 80’s, and 90’s respectively. SOA is a powerful concept and the details will get fleshed out as the implementers experience and solve the many problems that undoubtedly await. Yet even today the broad outlines of the business case for SOA are clear. These are:
I’ll talk about each of these briefly, but the point is that just as in the earlier schema for IT architectures (mainframe, client-server, etc.) the key issue is efficient resource allocation and utilization. Unlike the earlier schema, whose end goal was an essentially static infrastructure SOA implies a dynamic IT environment where automated service provisioning and IT process automation enable application services to wax and wane with the rhythms of business requirements. Rather than over-provisioning to meet every conceivable application requirement, UC offers a model of a flexible infrastructure readily re-provisioned for changing business needs. So at its core UC is about dynamic allocation of resources, not a static re-modeling of existing resources. And the best method for the efficient resource allocation is a market.
Part II will appear within the next week. In the meantime, comments welcome.
In the Cnet article Gartner: Prepare for consumer-led IT about a recent Gartner Symposium, the world’s largest IT consultantcy sounded the alarm for the F1000 glass house attendees:
. . . Gartner’s director of global research, Peter Sondergaard, warned conference attendees that consumerization will be the most significant trend to have an impact on IT over the next 10 years.
“We stand at the foot of a new high tide,” Sondergaard said. “There is a shift in technology ownership.” . . .
“Consumers are rapidly creating personal IT architectures capable of running corporate-style IT architectures,” he said. “They have faster processors, more storage and more bandwidth.”
He advised corporate IT executives to adapt to the changes and prepare for what he called “digital natives,” or people so fully immersed in digital culture that they are unconcerned about the effects of their technology choices on the organizations that employ them.
I’m glad he got the word storage up front, for the rest of the article, though perhaps not the actual Symposium, went on about mashups, AJAX, wikis and the like until I was ready to gag. These are all wonderful things, to be sure, yet unless the infrastructure supports the promiscuous and creative waste of storage, bandwidth and CPU cycles implicit in consumer-driven infrastructure, there can be no consumer revolution.
Mass customization, with bits
IT culture is rooted in high-volume manufacturing, where eliminating variables and consistent delivery are paramount. In contrast, today’s social-networking web app mashups are a drunken Saturday night in a mining camp. It is all flaming individuality and “look at me” capers with the occasional gun - or blog - fight.
When Novell was growing like a weed in the ’80s, IT wasn’t buying it. The corporate divisions were, because they’d bought PCs and wanted them to talk to each other. IT guys smirked at Netware’s lousy availability and primitive services as they tried, and failed, to get those PCs linked into the corporate IBM Systems Network Architecture hairball.
Don’t get me wrong, SNA worked well. It was simply costly, inflexible, closed and architected for a bygone era. Spending almost as much to interface to SNA as the PC cost was a very hard sale. Does that sound a little like today’s FC SANS? It should.
Industrial vs consumer
Industrial production is based on volume efficiencies or economies of scale. Broad-based consumption, or consumerization, is based the use of industrial technology to create small wasteful parcels whose advertising, packaging and transportation costs far outweigh the cost of the product inside the box. So it is with a consumer-driven infrastructure.
Consumers will keep things on disk, if they can, for decades (I do). They will email a 9 MB file to the next room rather than put it on a thumb drive and walk it over (I do). Consumers will download dozens of webpages to find a couple worth reading (I do).
Super-size that infrastructure for you?
Gartner calls them “digital natives”, but to glass-house civilization they look like barbarians at the gates, a mob of digital zombies mindlessly consuming everything in their path.
The technologies for cheap, high-performance, high-capacity storage exist. The demand is gathering strength. Yet as obvious as these trends are, it still looks like IT will again barricade the glass house to keep them out. The IT vendors who wait until IT surrenders will have waited too long.
Everyone else is writing about GooTube today. Yet I think my thoughts are original enough to be worth reading.
Another storage solution wrapped in an easy-to-use application, like iTunes, YouTube’s founders and venture capitalists have struck it big. Each of the founders will get about $600 million, while the VC gets a 40x return on its one year old investment. That is a home run.
More importantly, Google’s $1.65 billion purchase of on-line storage provider YouTube - iTunes in reverse video - is a concise lesson in the value of marketing.
Failure carries a price
I admire Google’s tech chops, as I’ve documented in several posts while I’ve been equally dismayed by their inability to gain market share in anything outside of search.
Google is an advertising company and the world leader in monetizing internet eyeballs. They also do a terrible job marketing their services. Google claims its Google Video product will live on, but with their brand recognition the real question is: why didn’t Google Video win?
The basic problem is that evidently no one told the engineers that the site had to be dirt simple to use. No finicky formats or video conversions for customers to get frustrated by. Just load and go. Like the Google homepage. Marketing is supposed to know things like that. Clearly, no one is minding the marketing store at Google.
The beginning of the end of Google’s culture
In its 21-month life, YouTube has created value at the rate of $80 million dollars a month, sending every venture capitalist and smart kid searching for the “next big thing” in Web 2.0.
Founder Chad Hurley has proved that even a BA in Fine Arts can end up a billionaire.
If Google can fail at video, what other failures lurk?
Google’s culture will suffer from the YouTube acquisition. Sure, the guys on the Google video team will wondering why they aren’t richer. But the bigger hit comes from knowing that YouTube isn’t about the technology, it is about the market. And there Google didn’t have a clue.
Every Google engineer has to be wondering if their next project will be superseded by something smarter hatched outside the company. And that also makes their founders culture heroes and insanely rich. Sure, Google is still a great place to work, but if it is no longer hatching the Next Big Thing, some of the top talent will get restless. As soon as they vest they’ll be gone.
I don’t think Google will ever get its marketing act together and reach their full potential. Nor is it clear that they will be able to maintain YouTube’s market dominance. Yet it is nice to know that YouTube is now backed by some of the deepest pockets in the world, protecting, I hope, the Fair Use provisions of the copyright laws.
I signed up for the ZFS discussion group over on OpenSolaris.org and have been following a fascinating thread about a versioning file system.
The discussion has been hot for several days. There is intelligent conversation about data preservation: CDP, snapshots, source control systems and file versioning among others. This is how features often get hammered out by engineering. The good news is that these guys are trying to grok the user experience.
File versions: the prequel
You know how everytime you edit a file in Windows, OS X or, I’m guessing, Linux, your saved file is now the only version of the file you have? It doesn’t have to be that way.
You could have the file create a new version of itself every time you edit it. So later, if you decide that fourth 321 Margarita wasn’t a good idea - it usually isn’t - you could go back to an earlier version, and begin again.
File versioning was standard on DEC’s VMS and TOPS-20 operating systems, and I really liked it. I could do radical things, secure in the knowledge that I could go home again.
ZFS file versioning
It doesn’t exist today, but the engineers are talking about. Here’s part of my favorite post so far, to give you a taste of the debate between two of the folks. I’ve put their comments into different typefaces so you can tell them apart:
Given this, we’re back into the problem FV is supposed to solve. It is entirely possible for an editor to keep open a file for a long time, periodically writing out your changes without issuing a new open().
You describe this as a problem, but *I* see it as the exact thing that makes file versioning useful. It DOESN’T save random magically chosen moments; it saves exactly all the version that *you*, the user, saved at some point of the editing session.
Word with auto-save turned off is a prime example. Given this, you’ve only created a new version when you first load the document, and all your intermediary changes are lost, since it only saves the document on close().
You’re forgetting that the user, unless he’s stupid, will save regularly during the editing session.
Thus, in order to get benefits from FV, your editor must issue periodic close() and open() commands on the same file, as you edit, all without your intervention. Exactly how many editors do this? I have no idea. So, the only way to enable FV is to require the user to periodically push the “Save” button. Which is how much more different than the current situation?
It is completely and utterly different from the current situation. In the current situation, when I type the “save” command *I am deleting a previous version*. That’s dangerous, because people don’t think of it as performing a destructive operation, and hence don’t give it the care and consideration they give to an explicit “rm”. And that’s precisely what file versioning fixes; saving a file is no longer a destructive operation.
The second guy has the better argument
Not being an engineer I’m not sure how to enter the discussion, except to say I think versioning was a great idea 30 years ago and I still think it is a great idea today.
Actually, it is a Gear6 StorageMojo.com podcast which stars moi.
This isn’t a happytalk Gear6 commercial. In fact we don’t say one word about Gear6 as I still have no idea what they are planning. You’ll get my opinion on Gear6’s product when they go public later this month.
I’ve been wondering if I should publish podcasts and have been unclear whether it would really work for people. If you listen to it please comment here and let me know what you think.
Fear mongering over the new Federal Rules of Civil Procedure (FRCP) requirements for electronic discovery has already begun with an article Storage Goes to Law School. As December 1, 2006 implementation date approaches expect the hype to rise. The “buy my widget or go to jail” line is well-nigh irresistible to sales and marketing folks, but way overstates the case. I’ve looked at the rule changes in some depth and they just aren’t that difficult. It is easy to fear the unknown. So let me give you the lay of the land. In less than 3 minutes.
FRCP: flow control for lawyers
The FRCP “govern the conduct of all civil actions brought in Federal district courts”. Federal district court judges don’t like their time wasted, so these rules are designed to ensure that by the time people get to court they are ready to rumble. These rules are binding only on the Federal courts, but many states model their rules on the FRCP, so expect to see these changes reflected in many state rules over time.
Because the Supreme Court says so, that’s why
The FRCP is produced by the federal judiciary as authorized by Congress. The rules process is long and involved with ample time for public comment and several review layers. Finally the Supreme Court approves the change and unless Congress intervenes, the rule becomes binding. It usually takes 2-3 years to change a rule, but the electronic discovery process changes have taken almost 6 years. They took their time and the result is, IMHO, fair and reasonable.
The nitty-gritty
Most of these rules changes are mostly technical (legally-speaking) in nature. Quoting from the uscourts.gov summary
OK, stop yawning, here’s the good stuff
The drafting committee’s report on the rule changes called out a couple of interesting points.
The StorageMojo.com take
The new FRCP rules are not, primarily, an IT problem. Your company’s record managers and legal advisors are responsible for figuring out company policy on record retention. IT should state clearly that they are ready and willing to help formulate cost-effective means of supporting discovery requirements, but under no circumstances should IT take responsibility for defining those standards. By this process IT will have the support of the firm’s legal advisors if any expenditures are required.
These policies will have to be documented and IT folks trained on them. You’ll need know that when the VP of sales storms in, demanding that his emails to a competitor be deleted right now! that you are on safe ground refusing until the proper process has been followed.
Now for something completely different:
The Illustrated Guide to RAID on one webpage.
In the convoluted way of the web Chris Mellor, of the UK’s Techworld site, got this from John Toigo’s site. I saw it at Techworld.
Infrastructure: it’s all pipes, pumps and pools. And the software to make it play.
Update - And check out the DVD rewinder here. Yes, you really can buy it.
The customer disconnect between what they need - I/O - and what they buy - gibabytes - is boiling over in the storage world. Hitachi Data Systems’ always thoughtful Hu Yoshida had a couple of posts whose juxtaposition crystalized the problem. The first post noted, in reference to Web 2.0 companies and the disk drive’s 50th birthdary, that
A great debt is also owed to the engineering innovations . . . which drove the cost of random access storage from about $50,000/MB to less than $0.002/MB. Storage is now cheap enough to give away freely and still cover costs and make a profit based on the services they provide from the stored content. [emphasis added]
Then, in an earlier post on industry revenues, he talks about storage consolidation:
. . . the growth rates for external disk storage have increased to about 60% over the last two years, while utilization of storage has been dropping to about 30%. . . . The CIO of a large financial company with over 2PB of storage said that his storage was only 20% utilized, and over 70% of it was expensive tier 1 storage.
Expensive, underutilized asset or cheap service?
Both observations are true. Both are in dire conflict. Yet it is a weak mind that can’t hold two opposing ideas at once. So what does it mean?
Two illusions
Economists refer to the money illusion, a term for an individual’s belief that a larger number of dollars means being richer, even if inflation is eroding the value of those dollars faster than the value increases. Individuals look at the number of dollars they have, not their purchasing power, and they get fooled thinking are better off even as their purchasing power declines.
In storage, the capacity illusion reigns supreme. We measure storage utilization by looking at capacity in gigabytes, which, as Hu points out, is the cheapest part of storage. The expensive storage component is I/O. And the expensive management component is people.
Run the numbers
Five years ago, the average disk drive cost about $4/GB while the average cost of OLTP tpmc was about $20. Today, 3.5″ disks are about $0.30/GB and OLTP tpmc is about $4. So capacity is less than 1/10th the cost of five years ago, while I/O is about 1/5th the cost. The relative cost of I/O has doubled in the last five years.
Go 1,000 miles between fillups!
The bursty nature of most I/O means that storage systems have to be over-configured to meet peak traffic needs. So the complaints Hu hears from customers have an underlying cause: arrays are configured to meet I/O requirements, but customers buy gigabytes, not I/Os.
It is as if people bought automobiles based on how far they go on a tank of gas. Everyone could tout a bigger gas tank, but the real issue is mileage. Likewise, everyone touts capacity, but I/O is the issue.
It is past time for vendors to change the conversation from gigabytes to IOPS. This would help the big iron vendors sell more big machines, while focusing customers on the I/O demands of their applications. It isn’t the whole solution, but it would be a good first step.
Comments welcome, as always. And I figured out how to turn off registration, so while comment spam is way up, I hope comments will be also.
Physicist Niels Bohr said that. But you don’t have to be a Nobel prize-winning genius to know how true it is.
Marketing people don’t try to screw up new products, because it is so easy even when you are doing your level best. It is just really hard to build a model of something as chaotic as a market, feed in the secular and technology trends, and then figure out what a product should look like in 18-24 months. And if the project slips a year even really smart marketers will end up looking dumb.
Want better products?
Sure, some vendors are just caught up in maintaining their current gig. Yet many vendors are also looking for the Next Big Thing, trying to build really cool products that lots of people will like and and buy. Which is a Very Good Thing.
A big problem is knowing when key elements of the market will be in place to make the Next Big Thing the Next Profitable Thing. Back in the ’80s lots of smart folks thought that bubble memory was going to kill the disk drive market. Oops. So even smart people have a hard time figuring out what is going to happen and when.
Dear Buddha, I’d like a pony and a tiny disk array
Take the transition to 2.5″ drives from 3.5″ drives. Much smaller, lighter and power-efficient than 3.5″ drives, the only thing holding back this transition is the excessive price-per-gig premium charged for 2.5″ drives. If a vendor knew when there’d be price parity between the two sizes, they could start engineering products using the new form factor well in advance.
The crystal ball. . .
Yet how can a vendor know this until it happens? Well, it turns out that markets have remarkable predictive powers that can provide good answers to such questions. Which is just what Storage Markets is doing.
Be a stock picker
Storage Markets issues “stocks” that embody different answers to a question of interest. You buy and sell these stocks based on your beliefs about the question. The stock price rises or falls depending on good old-fashioned supply and demand. If you predict well, the value of your portfolio increases. Storage Markets tracks all this for you, as well as listing the top traders.
How do they make their money?
Simple. SM charges to ask questions. To a vendor it is a market research expense. For you, it is a free opportunity to document your predictive Mojo.
Influence storage vendor investment decisions. Have some fun. Every IT department should have at least one designated investor. That’s right - put it in your job description.
Kudos to the Storage Markets team. Comments, as always, welcome.
