StorageRap

The many convoluted "benefits" of meta LUN management

NOTE  This post has been updated after being initially published. The EMC Storage Info blog posting was published a year ago on March 27, 2008 - not on March 27, 2009. Not only that, but the link to the post was bad too.  Arrggh.  My apologies for any confusion this may have caused.

---------------------------------------------------------------------------------------------------

The EMC Storage Info blog had an interesting post last week  from a year ago on how to accomplish wide striping with meta LUNS.  Here's a graphic that was published with the blog:

 
Even with concatenated Meta LUNs, this approach is a messy management problem that is unnecessarily complicated and gets worse over time as space gets used up inside the system. One of my fave cartoons, Pearls Before Swine, today made me think of meta LUN provisioning:

Compare this with autonomic wide striping in 3PAR InServ arrays (of course using thin provisioning)

1. Create a provisioning group (usually all the drives in a tier, where tiers are made with different drive types)
2. Then create volumes that accesses storage, including what RAID level to use
3. The array automatically creates a minimally sized volume, with micro-RAID sets spread over all the drives.
4. The array adds incremental capacity, as needed, from available storage across all the drives in the tier.

Now that VMware has endorsed thin provisioning and everybody has moved past the paranoia stage of this technology, it clarifies the implementation differences of enterprise storage. The combination of thin provisioning, mico-RAID and micro-sparing on chunklets and autonomic wide striping makes managing 3PAR InServ arrays much easier than anything built from meta LUNs.

Steve Duplessie's independent reality check on SSDs

3PAR customers really like our automated management.  The last thing they want is something that is a kludge to manage. They don't want to deal with stupidly complex provisioning such as meta LUN hashups.  The fact is, you don't need SSDs to achieve excellent mixed workload performance. 

The latest blog post from Steve Duplessie goes right at the management burden of SSDs. Here's a quote:

I would argue that stuffing Flash into that 1 file server/array/gizmo will cause MORE work to be done operationally – because now you have to move and migrate things to take advantage of that new sub-component – and every time you change anything, you open yourself up to new SYSTEMIC issues. Data/System migrations for capacity optimization or load balancing is a never-ending tactical nightmare of IT operations. Flash maybe kept someone sane by masking the need to do one of these migrations on that system, but only for a little while. In reality, it probably causes more work to leverage the benefits of the flash than it's worth.

If you think Steve is onto something then look at 3PAR's T and F class systems.  Great scalability, high capacity utilization, autonomic wide striping and superb mixed workload performance.

EqualLogic is a bright spot for Dell

Dell announced earnings yesterday that reflected the rough economy. One of their most obvious bright spots was EqualLogic, which according to Dell CFO Brian Gladden saw a revenue increase of 71%, although storage revenue overall was down 17%.  Although not mentioned specifically, I suspect this means their Clariion business with EMC was down too.

FWIW, I don't think this is a case where EqualLogic systems are displacing Clariion sales - although some of that is probably going on - but is instead a case where EqualLogic storage is successful in SMB accounts just as it always has been.

3PAR and EqualLogic arrays are similar in the level of automation they provide, which saves storage administrators lots of time and effort and is why both are popular now.  But other than that, the products are very different and developed for distinctly different types of customers, with some overlap in the medium sized business segment.

The virtual land grab favors cloud computing

Rich Miller at Data Center Knowledge posted this morning on how data center space is filling up.

This is another pressure point that will drive IT operations to cloud/utility service providers.  As data center real estate becomes increasingly difficult to manage, why not turn that problem over to a cloud service provider that takes care of all those problems for you while also providing SLAs?  Do you really want the headaches of playing resource tetris with multiple hosting sites because you can't get space in your current site(s)?

A cloud service that virtualizes all resources and guarantees service levels will look very attractive compared to hosting moves and migrations.

The end of free love at VMworld? Ok

VMworld revived the IT conference landscape by being open and energetic.  I remember being totally stoked at VMworld a few years ago in LA by the enthusiasm nearly everybody had for the event - vendors and customers alike.

Brian Madden has interpreted this years VMWorld sponsor and exhibitor agreement as being much less open in the past by not allowing demonstrations of competing products. The comment thread on his blog will likely grow by leaps and bounds today, so people interested in this discussion might want to link over to his blog and follow it.  Scott Lowe has also posted on this on his blog.

So far, most of the reaction has been to ask VMware how they could do this and to point fingers at Microsoft for having closed conferences.  Some people will do anything to take a swing at Microsoft. If VMware is doing this in reaction to anything Microsoft does with its conferences, then I'd say their motivation is backwards.

The great love-fest of VMworld has served VMware very well. The show helped create a community - not just for VMware - but for the entire virtualization industry. The first job in any new industry is to create a market and VMware has certainly done that.  Having done that, VMware has moved on to optimizing VMworld for its business opportunities - and that means for the ecosystem that is aligned with it's technologies.  VMware is not in the conference business and it needs to focus on its big and challenging business agenda, not on producing a love fest for virtualization. This is an opportunity for conference producers.

I wouldn't expect EMC to open up EMC World to 3PAR and other vendors that compete vigorously with them.  Why would they go to the effort of producing an event that would risk losing customers to other vendors? That would be insane.  I don't see why VMware should be any different.  

What's more interesting to you - Hitachi's TVs or their clustered storage?

Barry Whyte from IBM probably wishes he could re-write his blog post for today that anticipated something big from HDS, because HDS underwhelmed the storage world with it's announcement of storage pork - otherwise known as HAM (High Availability Manager).

I think Hitachi was trying to figure out how to get some buzz going without buying a hot startup, hiring an expensive exec from a competitor, laying people off or cutting everybody's salaries. Nobody ever accused Hitachi of marketing originality, but trying to create blog buzz for software that eliminates a single point of failure from enterprise products might place them in the top percentile for marketing ineptness. 

People were expecting a lot more and maybe sometime there will be more - about Hitachi clustered storage.  3PAR has had it since day one - and we celebrated our 10 year anniversary this month.  I know that Hitachi would love to change the world, with their cluster attempt, but they are 10 years after, after all.

This clustered storage stuff is hard.  Its easy to talk about, but much more difficult to deliver.  That's why 3PAR's track record of delivering clustered storage arrays means something. 

Otherwise Hitachi announced new big screen TV products today, something that might be interesting to blog readers. While their USP storage arrays tends to be expensive, their new Alpha line of LCD TV's look like they might be competitively priced.

VMware gets a piece of Terremark's rapid rise

I just saw that VMware had invested in fast-rising cloud service provider - and 3PAR customer, Terremark.  I think this is an astute move by VMware, both as an economic investment in an early market leader, but also as a means to increase the strategic technology relationship between the two companies.

Terremark's Enterprise Cloud could is an excellent example of what the future of cloud, or utility computing will be. People who think cloud computing is just the next great forgotten buzzword should take a look at what Terremark is delivering.  Its a way for their customers to run their applications using only the compute, storage and network resources they need.  It is the model of information infrastructure efficiency 

Foskett forecasts Google lame-search apocalypse

My friend Stephen Foskett is a brilliant guy who seemingly never stops thinking about technology, except when attending Little League games.  Ont Friday he posted on how Google searches are destined to lose their value over time by not including the most relevant links.  Acronym junkies might enjoy talking about this de-evolution of search as SANAS (Search as Noise and Spam).  To me it seems like another example of the tragedy of the commons - like the space around our planet.

So who loses if Google search start sucking?  Certainly not Google. Behavior is hard to change and as long as Google is free to use, people are going to search it to smithereens (SITS).  Many techdroids have developed a knee jerk reaction to curiosity, and it's spelled Google!  And once we go there, its all about the distractions.

This post marks the return of the SWCSA, which has been on hiatus for a couple months.  There is something about the freedom of Summer that just makes me want to start rambling into the windshield.

Here are links to blogs referenced in this video:

First ever steering wheel conference video

Autonomic Wide Striping Illustrated

Looking for a beautiful Asian American

To save another.  Nick Glasgow is a guy who works in our industry who needs a bone marrow transplant.

The search for a donor for Nick is "on" on the Internet - so if you fit the profile and can be a donor to Nick, you will definitely be known as one of the most beautiful people on planet earth.

If you don't fit the profile and you can blog or tweet about it, keep this search alive.  With luck, we'll find somebody somewhere.  Use the image and link below in your blog.

Netapp couldn't beat Data Domain, so it eats them instead

The storage industry screws just keep turning.  A few minutes ago I heard that Netapp was acquiring Data Domain.  Congratulations to the stockholders of Data Domain, but I hate to see another growing, independent storage company get swallowed.  Once Data Domain joins Netapp, 3PAR and Compellent are going to be the only publicly traded  array vendors with consistent sales growth over the last couple years. 

This deal surprises me because Netapp's VTL products compete with Data Domain's products. Last October Chris Mellor from The Register wrote on how Netapp finally had an answer to Data Domain, Quantum and EMC.  Apparently that answer needed to be amended six months later with cash and stock worth $1.5B. Netapp will have to spin this carefully to their shareholders and VTL customers. It certainly is possible to have overlapping products in the product line, but I'm not sure how well that will work for the de-dupe market. 

One part of the story could be a SAN-only storage array from Data Domain - meaning it wouldn't come with a NAS head. This would give Netapp access to the SAN market in a way they have never been able to do with their Filer products. 

I can't help but wonder if part of this deal is about generating interest in Netapp.  Lately EMC, 3PAR, Compellent and Sun have been generating most of the attention in the storage industry and Netapp has been pretty quiet.  That will change for a few months now (or at least until the next big acquisition occurs).

Otherwise, Data Domain lacks a corporate rapper, but if you ask me, they never had that far to go to catch up with Netapp. Still 3PAR's 3P might have to step it up a notch now that the Double Team can tag team with the Netapp rap crue.

HP: So much for that pay cut excuse to avoid layoffs

News last night that HP plans to cut more workers after announcing in February that they were going to cut everybody's salaries to avoid layoffs. Seems like HP has done the old bait and switch to their employees with this one.

HP employees are probably not real motivated by the news, if the comments on this Barron's article are any indication. OK everybody, work hard and make less!  Repeat!

This is exactly the sort of thing that I've been concerned about in this economy.  Wages being cut by companies with lots of money in the bank that are making profits. Yes, I know the responsibilities of management lie with shareholders, but does that mean that employees don't matter?  In the end, customer service takes it on the chin and that is very difficult to repair.  Does that then mean that customers don't matter either?  

Autonomic Wide Striping Illustrated

3PAR's autonomic wide striping is a technology that is somewhat counter-intuitive and sometimes people have difficulty understanding how it works. So, I made this video, complete with stop-motion animation to help people visualize the differences between it and traditional array volume management.

The data-to-disk layout in 3PAR InServ arrays is determined by algorithm, as opposed to being manually configured as it is with most other arrays.  This saves administrators many days of resource planning and saves companies from needing to spend money on professional services to perform a task that lends itself so well to computerized automation.

Newest Infosmack podcast is online

***  Updated May 20, 2009 **  Infosmack podcasts are now available on iTunes here.  If you have iTunes installed on your computer, click the link to open the Infosmack page on iTunes.

Greg Knieriemen and I host a weekly podcast we call "Infosmack"  This week's Infosmack features John Webster from Illuminata, Inc and Storagezilla from EMC.

Infosmack covers a wide variety of data center topics, with a lot of talk about the storage world.  This week's show covers Cisco/EMC acquisition rumors, Oracle/Sun's potential, Netapp, Brocade, Twitter and the new Star Trek movie.  Click the player below to check out a few sound bytes from the show.  If you want to hear more, click the link above for this week's episode.

Powered by Podbean.com

 

Steve Duplessie on the dynamics of cloud computing

Steve Duplessie from Enterprise Strategy Group published a report recently giving his impressions of the role cloud computing will have on the future. I don't want to spoil it for you by saying too much about his thinking, but I thought I'd share his persepctives on the increasing importance of business agility:

As Web 2.0 migrates from a world of questionable business value to mainstream business deliverables, IT departments are increasingly finding themselves without a plan. How does an IT operation go from taking three months to plan and provision infrastructure to support a business application to creating and scaling one in semi real-time? Without any realistic method to determine what the extent of interactions will ultimately be or how fast growth will occur, it is very difficult to make any plans.

Definitely a good piece that helps clarify "cloud".

Google's outage and real cloud computing

Google apparently is having problems with gmail access today as CNET reports.  Anytime this sort of thing happens, questions come up about whether or not cloud computing is ready for prime time.

The answer is - of course it is, but maybe not at Google. (FWIW, my gmail account is working fine and I haven't noticed any problems.)

Yesterday, Dan Chu from VMware responded on his blog to a recent Google blog trumpeting Google's cloud computing approach.  I think both of these are good reads, but at the end of the day Dan's arguments about application support in VMware-based clouds clinch it for me.  Why would you want to adopt a cloud infrastructure if it meant changing the way your end users work?  Here is a quote:

Customers are looking to match their IT platform to their business needs, not the inverse. The Google approach calls for a least common denominator set of non integrated cloud services that everyone squeezes into. Customers want the flexibility and breadth of solutions that exist today along with the efficiency of the cloud.   Customers are not about to re-write or modify their applications so that they can run in a specific cloud.  In particular, given current macro-economic circumstances, customers have a high priority for a cloud platform that can take their existing apps, and enable them to take advantage of the cloud. 

J Schwan left an interesting comment on Dan's blog, citing the leverage of multi-tenancy in public cloud infrastructures and what he sees as the eventual economic end game of these types of clouds.  Multi-tenancy is certainly an important advantage for cloud service providers and is why 3PAR has virtual domain management technology in our InServ arrays.

Dan mentions VMware is used in "more than 500 service providers including major global players like  AT&T, Savvis and Terremark".   FWIW, AT&T, Savvis and Terremark all use 3PAR storage as part of their cloud service offerings.

Another example of real-world cloud computing is our customer Attenda (in the UK), which has been recognized as VMware's EMEA Hosting Partner of the Year for 2008 and 2009.   They recently announced that Regus, a global workspace services company has decided to run their line of business applications on Attenda's utility computing services. 

The ability to run line of business applications on cloud services is already happening at Attenda and other cloud service providers.  These visionary companies are certainly not waiting for Google to figure it out - the question is whether Google will ever be able to catch up, if ever. The train has already left the station.