Tiered, Schmeared, Weird: Enterprise Storage Management Issues and Technologies

We've been doing research in the areas of enterprise storage management with a focus on what is now called Tiered Storage and EMC calls Information Lifecycle Management (ILM). The technology was also called, back in the day, Hierarchical Storage Management (HSM).

The basic idea is pretty obvious. Store the most critical, performance sensitive data on devices that are the most reliable and highest performing (and most expensive). Store less critical data on less expensive devices and the least critical data on the least expensive and possible off line devices.

I'll probably address several of the aspects of these technologies in multiple postings over the next several weeks. For now, I'll just define some terms and establish a baseline context for the future. For these discussions most of the storage solutions mentioned will be shared. Sharing may be on a SAN, NAS or iSCSI protocol, but the devices themselves are shared.

First of all, lets describe the nature of the hardware typical to the different storage tiers. As mentioned previously, the logic behind tiered storage is that the most performance sensitive, mission critical data should be stored on devices that offer the greatest reliability and performance. This generally means Fibre Channel Storage Area Networks with the largest arrays, fasted backplanes and connections. Vendors for these systems include IBM DS6000 and DS 8000, HP XP series, SUN StorageTek 9900 series, EMC Symmetrix, Hitachi Universal Storage Platform, Pillar Axiom, and a few others.

The second storage tier, at least for purposes of this discussion, consists of systems that are not as powerful as the top of the line products. They are intended for servers and applications where maximum performance is not the key driver. The hardware can be the same as previously, listed, perhaps with lesser components, or be older models. Depending on the scale of the organization, these devices could also be the same vendor's mid-range systems. For example, HP EVA or EMC CLARiiON. This tier also opens the door for additional vendors and connection protocols. For example, NetApp filers. (I know that NetApp makes high end devices and could be included in the previous list, but this discussion is already very complicated.)

The next tiers bring us to another aspect of the story: online, near-line and offline storage. Quickly, online storage is that which is connected to and immediately available to the devices needing the data. Near-line is storage that's connected, but which may not be immediately available for use, such as a Magstar tape used in legacy mainframe systems. Offline storage is data that needs to be brought online before it can be accessed. This can be as sophisticated as a DVD stored in an automated library or as simple as a tape stored in a vault. The tiers being discussed now can consist of any combination of devices that fit into these three states depending upon need and design.

My next post will discuss some of the ways vendors are attempting to both provide tiered storage solutions. In the future, I'll be looking at solutions from vendors who purport to provide alternatives that greatly simplify this architecture such as XIV Nextra, and perhaps drill down into some of the technical details. I'll also be discussing how the various vendors go to market with their interpretations of these themes. Not to belabor the obvious, but each vendors strategy exploits their history, market and product strengths.

The Server Virtualization Battle

Citrix's Xen won't cede to VMware or Hyper-V

It seems like only yesterday when I was wondering about whether the price war in the hypervisor market should be worrying VMware. Wait, it was yesterday. Today I read that Citrix won't concede in the price war or on the technology.

From the TechTarget web site:

In December 2006, two startups, XenSource Inc. (now owned by Citrix) and Virtual Iron Software Inc., kicked off a virtualization price war, offering virtualization for as little as one-sixth the cost of VMware. Last fall with the entrance of Oracle Corp., Novell Inc., Red Hat Inc. and others into the battle, the price competition intensified, and then this spring, rivalry flared when Citrix cut prices again and initiated flat pricing for servers with up to four sockets. Citrix's efforts have met some success as well. Now all the players have geared up for Microsoft's August 2008 launch of Hyper-V, which is extremely low cost at the price of $28 per server. Citrix has a special mission in this new lanscape. Sandwiched in between feature-rich VMware and lower-cost Hyper-V, Citrix's Xen has the daunting task of remaining price-competitive yet fully featured enough to compete.

Citrix's Xen won't cede to VMware or Hyper-V

Yesterday was about VMware not lowering prices because they "…have the right product packages with prices that allow the customer to do whatever it is they need to do with our technology…".

Today is about how "…Citrix may ultimately focus on the desktop virtualization. First, Microsoft hasn't targeted its virtualization efforts to virtual desktops, leaving an opportunity for others to gain advantage."

And that Citrix "…has an open storage interface that integrates innovations from storage vendors into the hypervisor with plug-in drivers."

The rest of the article combines some market-speak with lots of assertions about feature comparisons between Citrix XenSource, VMware and even Microsoft Hyper-V (though to be fair, they state clearly that since Hyper-V isn't shipping feature comparisons are invalid at this time). Lots of on-the-one-hand vs. on-the-other-hand comparisons.

Actually comparing these products and either publishing the results for all to see or sharing them privately with our vendor clients is what my company does best. VMware: hear those footsteps behind you? Novell: wonder why you're not in the top tier? Virtual Iron: the reviewers like you, but don't you need help getting some real market share? Microsoft: we know you're going to deliver something functional eventually; wouldn't it be good to use have some support for your product claims and debunk the critics?

Why don't you give me a call? 

VMware won't lower price, says exec. Does pride go before a fall?

The link will take you to the whole interview, but the point is made in the headline.

VMware won't lower price, says exec

The executive, senior director of product marketing Bogomil Balkansky goes on to explain,

"...The objective is to make sure that we have the right products at the right prices for all types of customers. I don't think that principle is going to change.

I would argue that right now, we have the right product packages with prices that allow the customer to do whatever it is they need to do with our technology. Whether it's something simple like partitioning a server for test and development, a single server consolidation, production server consolidation for a larger environment or if they want to reap the other benefits of virtualization, such as business continuity, disaster recovery and dynamic resource management, they can do that at a price that affords a tremendous ROI. Virtualization is one of those technologies where the ROI is quick, undeniable and easy to see. "

The interview then goes on to address the competitive challenges that VMware is facing, denying that the competitors will have any immediate impact on VMware and in that though Microsoft will become a formidable competitor, they're years away from being even close to VMware in delivering a server virtualization platform with the management functionality enterprises expect.

Mr. Balkansky makes very good points about the strengths VMware brings to the table, and is essentially correct in his assessment of the current server virtualization landscape. The viewpoint is in most essentials that offered by any market leader who has also created the market. In marketing, being first to market with a successful product and sticking with it until the product gets accepted and becomes a de facto standard is the place to be.

Our analysis of server virtualization, including the early work we did in developing tile based virtualization capacity benchmarks, has all centered on VMware. Our research on competing hypervisor and virtualization management not only has looked to VMware as the leader, but has recognized the overall superiority of VMware's products.

There's just this one thing: market leader arrogance. As I said, the viewpoint expressed in the interview is that offered by other market leaders. I'm concerned though about the arrogant tone: no one out there is a real competitor, so we're not really worried about them. We'll watch Microsoft, but Citrix? Xen? non-issues.

Nothing wrong with the assertions. The facts bear them out. But Citrix release of XenDesktop will have an impact, and just because Microsoft hasn't delivered a full management and deployment model for Server 2008 with Hyper-V doesn't mean the immanent release won't be important in the marketplace.

Mr. Balansky may not be willing to express concern in an interview (though I'd be surprise if the filings for the stock market will be so sanguine), but he should be looking over his shoulder. A free, fairly strong offering from Microsoft will have tremendous impact. Especially in the smaller businesses where the costs for ESX and Infrastructure III are perceived as prohibitive. Sure, a company buying two or three or four VMware licenses is not to important VMware. Most vendors don't think about these small sales too often. But there are tens of thousands of companies with a half dozen or so under utilized X86 servers looking at virtualization. Microsoft's offering will be very attractive to these folks. The lack of a full suite of management tools will probably not matter too much since these companies are probably not using that much in the way of sophisticated management platforms anyway.

Hopefully, VMware is investing in following the success of its competitors and looking closely at how their products actually deliver rather than only at the market numbers. Otherwise, a year or two from now, they might be surprised to see their market share shrinking. And shrinking in their existing customer base, where it hurts.  

Apple and Business — Is it 25 Years Too Late?

Jason Perlow, writing on Larry Dignan's blog at ZDNet.com isn't impressed. (I think it's Jason as that's what it says on the blog, but I may be misinterpreting the page design.) The lede:

Hey, Apple’s releasing the Insanely Great iPhone SDK so we can all write enterprise iPhone applications! They’re going to make the iPhone compete with the BlackBerry and hook it up to corporate Exchange email servers! Whoopee! Apple has a Business strategy!

Apple and Business — Is it 25 Years Too Late? | Between the Lines | ZDNet.com

I think he protests too much.

Most of the points make are not in error, though obviously intended to incite fan boy rage. I've had similar thoughts myself for years. And expressed them to Apple executives and other industry folk.  I have only a few issues with the entry.

First of all, it's obviously written from a "anti-Apple" perspective. I know I've probably been reading this column for years, but I'm not a fan of anything and don't recall whether this blogger (either Larry or Jason) are pro-Apple, anti-Apple or just the sorts of writer who like to put down people and organizations because the negativity excites readers and builds traffic. Perhaps they're positioning themselves as the Rush Limbaugh of computing? My issue - enough with the Apple and Steve Jobs bashing. It seems you're jealous or something. Get over it.

What's more important and much more annoying is the assumption that Apple wants to have an enterprise computing strategy that makes their products just like everyone else's.

Some quotes to illustrate:

For Apple to have a real Enterprise strategy, they are going to have to do better than iPhones and sex appeal.

Much of this enterprise cluelessness stems from a 25-year history of Steve Jobs and his legacy at Apple of not really caring about what businesses actually want.

Mac OS X technology is elegant, and there is nothing wrong with it from a pure software architecture perspective. The problem is that Steve Jobs and Apple doesn’t really give a a damn about how to apply the technology to business and make it attractive to enterprises in order to mass adopt it.

The last sentence is the important one. If my interpretation is true - that the author, whomever it may be, believes that Apple's enterprise computing strategy is to conform to the mold set by IBM, Dell, HP and so forth and that mass adoption is the ultimate goal for any computer technology company selling to business - the author is missing the point. Apple doesn't need to be HP or Dell. It needs to be Apple.

Not all business users need to have absolute conformity. Or don't you actually recall the point of that ground breaking 1984 ad?

All Apple needs to do is be a good enough corporate citizen so that enterprise computing groups don't reject it out of hand and enable those would benefit from the enjoyment of an iPhone or the greater good looks,  stability, security and reliability of an iMac, MacBook Pro or whatever to participate as network peers.

Maybe I'm just an old hippy, but conformity is not the only, or even the best corporate culture to emulate. Even IBM doesn't have the corporate culture that that 1984 ad satirized.

I agree about some of the technical arguments and would love to see Mac OS servers running on enterprise class hardware at least as VMs on x86 boxes from various vendors on any virtualization architecture (VMware, XenSource, Hyper-V.)

It would be even more important if Apple collaborated with the BMCs, Tivoli's  and CA's of the world to make sure that enterprise management platforms can easily administer Macs without the absolute need to use Apple's own tools, excellent though those may be.

But go for dominance? Why? That's Microsoft's thing - and all that's done is get them in trouble they didn't need. If Apple can beat their 1% of smart phones target with the iPhone this year because of corporate acceptance that's great. If they can get even 5% of corporate laptops to be Macs - that's even better. And if the get secretarial desktops and executive desktops to even 5% it would be a huge jump in sales for them. Why do they need to strive for bigger things right now?

Quote for the day

 

There are grammatical errors even in his silence.
  - Stanislaw J. Lec

A Tragic Flaw

Watching the debate tonight...

There is a tragic flaw in our precious Constitution, and I don't know what can be done to fix it. This is it: Only nut cases want to be president.
  - Kurt Vonnegut

Installing Oracle on Linux is too difficult

A blogger at Tech Republic, Rex Baldazo, says that Oracle installation on Linux isn't as easy as it could be.

I still think Oracle could stand to spend more time making its Linux installation smoother.

Installing Oracle on Linux is too difficult | Programming and Development | TechRepublic.com

The author goes on to say that his problems may stem from being more familiar and comfortable with Windows:

"I’m neither a Linux admin nor an Oracle DBA, which partly explains why I’m having so much trouble getting Oracle installed properly on a Linux machine. But the installation is just too complicated, especially compared to the same installation on Windows."

The company I work for, Edison Group, inc. has performed installation of Oracle on Windows in comparison tests for several years now. One of the things we've discovered is that Oracle's Database was easier to install than even Microsoft SQL Server, especially SQL Server 2005. It made sense for us to install on Windows for these projects since comparing how SQL Server installs on Windows with how Oracle installs on Linux made no sense. We also compared Oracle Database 10g with several versions of IBM DB2 UDB. We used Windows for these comparisons too to maintain the same context as the MS SQL Server comparisons.

Maybe our premise was too limited for today's datacenter realities. It seems that when installing on Linux, the Oracle installer doesn't address pre-installation steps directly. According to Rex:

"Even when I thought I had the pre-install steps completed properly, the Universal Installer would complain about one thing or another. And, of course, the darned Installer doesn’t have any ability to fix the problem — you just have to quit the installation"

He compares this to what happens with SQL Server:

"...Microsoft’s SQL Server install, which can handle much of its own pre-requisite software if necessary.

...even if you follow all the pre-install steps properly and the Universal Installer runs without problems, you still don’t end up with what I would call a working installation. You still have to do post-installation steps manually to get the database to where it will automatically restart when the OS is rebooted."

It seems to me that Oracle has a challenge here as great as the one that instigated the original research we did for them: can it maintain the ease of use advantage that Edison has been measuring over the years on other platforms? After all, Oracle seems to use the same installer for all the platforms. Why isn't it as thorough on Linux as on Windows? Or does the installer work better on Oracle's Red Hat distribution where they have control of the bundled packages and perhaps access to RPM tools specifically for that distro? It wouldn't matter if the difference isn't that great since everyone in IT knows that installations on Linux are usually more complex than on Windows. Oracle has been pushing Linux as the preferred OS for their eponymous database server for several years now. I would think that they should be making every possible effort to insure that their product offered the same ease of use capabilities on Linux as on Windows.

Edison has been negotiating with Oracle to update our research to reflect new products from all the vendors. They have been reluctant to fund the additional research required to update the procedures that have been in use since 2000 to reflect the changes in the products. The issue raised here adds some additional areas of concern that Oracle should have Edison address. I'm not hopeful that Oracle would be willing to have Edison both update the procedures or look at Linux when comparing Oracle to IBM DB2 UDB, let alone Oracle on Linux to anything on Windows.

To me, that comparison, even if Oracle on Linux is not quite as easy as on Windows, would be a powerful competitive message. One that would leverage the ease of use message on top of the unbreakable Linux message.