Thoughtput

EMC Buys Iomega

EMC has finally closed the Iomega acquisition. Erick Schonfeld at TechCrunch has an insightful statement...

"Perhaps EMC could combine Iomega with Mozy, an online storage service it bought last year for $76 million. The trend in consumer storage, as with consumer software, is to offer both online and offline capabilities."

When one company owns everything from the external USB disk drive through online backup, and a suite of services for sharing in between, then we should be backing up more often, transferring data more often, and someone will be profiting more often.

One more example of EMC's current have-no-fear acquisition strategy.

Google I/O - great conference name!

As spotted on GigaOM, Google is hosting a conference for developers of web applications.

Google I/O is a developer gathering focused on pushing the boundaries   of web applications using Google and open web technologies. Google   engineers and web development leaders will lead you through two days   full of in-depth breakout sessions on the latest technologies, hands-on   Code Labs, and informal Q&A at Fireside Chats.

Now that is a great conference name! Click here for the conference website.

The amount of activity around web services and cloud computing recently is unbelievable. We had our first cloud computing roundtable recently and are planning more soon.

Bits of Memory and Virtualization News

Here are a few interesting tidbits that popped up in my email box this week.

Massive Memory Systems. Memory is the new disk! ....Disk will become the new tape, and will be used in the same way, as a sequential storage medium (streaming from disk is reasonably fast) rather than as a random-access medium (very slow). Tons of opportunities there to develop new products that can offer 10x-100x performance improvements over the existing ones.
click here for the full blog post from Computing at Scale

High-performance computing, mondo memory and new style applications
Just as storage and networking have been disaggregated from the computer, some amount of memory and processing, at least for specialized purposes, may also migrate on to the network
click here for the full blog post from Server Specs: A SearchDataCenter.com blog

AutoTrader plans VMware move from iSCSI to NFS
(more evidence of the NFS train leaving the VMware station...)
AutoTrader.com said it's getting ready to shift VMware data stores from the iSCSI side to the NFS side of its NetApp 3040 disk arrays, saying that the move will make VMware storage management easier and more efficient.
click here for the full story from Beth Pariseau at SearchStorage.com

Q&A on Virtualization for Everyone Blog

We had a chance to connect with Tarry Singh who maintains the blog Virtualization for Everyone. After our call, he sent us a list of interview questions.

Check out the full post here.

Haven't We Seen This Movie?

Stacey Higginbotham launched a post today on GigaOm titled Next Up: I/O Virtualization? discussing options to consolidate Ethernet, Fibre Channel, and InfiniBand connections on a single link. It is great to see companies tackling the path management mess created by trying to jam 6 PCI cards in a 1U server.

We spent some time discussing I/O virtualization and virtual I/Os with Greg Schulz at The StorageIO Group and have our own post and short podcast at Defining Virtual I/O Acceleration.

But more importantly, this continued activity in the I/O virtualization arena triggers the following...

How the heck did we go from:
1. We have Ethernet to
2. We have Ethernet and Fibre Channel to
3. We have Ethernet, Fibre Channel and InifiniBand to
4. We can put everything back on 10Gig Ethernet.

Is it just me, or does this seem like we could have saved a few billion dollars over the last few years and skipped a couple of steps?

I know there were things Fibre Channel could do before 10Gig Ethernet came around. And I'm sure someone will be happy to point out the interconnect efficiencies of InfiniBand. But let's be honest with each other...haven't we all seen this movie again and again?

The Flashtastic DMX-4

Today industry bellwether EMC announced that it will use flash-based drives in the newest Symmetrix DMX-4 arrays. While this might seem like a totally new concept, we should all keep in mind that EMC began life as a memory company, and due to a combination of their smarts and IBM's mis-steps, they turned their memory expertise into a storage system capable of siphoning off data (and revenue) from IBM's largest customers. This remains a fine example of the twists and turns of the technology industry and the need for all large companies to stay on their toes.

There are dozens of articles today talking about the finer points of the technology and where it fits. It will take a long time before the dust settles and a general consensus emerges on customer segments, applications, and workloads. In the meantime, it is fun to watch. See below for a few links.

And while many will focus on the performance gains that memory can deliver over traditional disk drives, something Gear6 has long discussed, I found this quote from the Wall Street Journal to be the most enlightening,

Recently we had to buy a lot more storage than we need in order to get better performance. This technology looks interesting.
-computer manager for a large U.S. financial institution commenting on EMC's announcement

So is this about performance, or is it about re-thinking the architectural model for storage? Where are the data centers that can maximize the benefits of memory and disk drives? The reality is that they are well established at Google, Yahoo!, Amazon, and other web giants. And they are soon coming to an enterprise near you.

Here are a few of today's posts:

EMC's New Flash Drivers - StorageMojo

Enterprise Storage Strikes Back - Chuck's Blog - plus lots of other links at the end of the post

Flash!  EMC’s DMX is the New New Thing Again - Stephen Foskett

and a vigorous discussion on the Byte and Switch message board - Personally, I'm amazed at the amount of this media vs. that media discussion on this message board, instead of questions like, "given that we have all these great technologies, what is the best way to apply them across the data center?"

Ending the Application Waiting Game

Enterprise Networks & Servers just published a piece we wrote on Ending the Application Waiting Game. Understandably there are a lot of reasons why applications might have slowdowns, but increasingly we see  delays in I/O access as the root cause.

The article also includes a section on Top Symptoms of Storage Bottlenecks, which should be helpful for those trying to determine the source of the problem.

Arthur Cole at IT Business Edge has posted a relevant blog piece on Improving Application Latency, which starts with the following excerpt:

Time was that when an application took a few more minutes than normal to boot up or failed to access needed data right away, it led to a few choice words by the user and that was that. But in today’s world of split-second financial transactions and real-time medical imaging, application and data latency can have serious repercussions. [read the full post here]

Gear6 Video from StorageMojo

Robin Harris at StorageMojo hosts a short video about Gear6 including an interview with our Chief Technology Officers, Nisha Talagala. The StorageMojo post is here, or you can view the embedded video below. Enjoy.

More on Google's Storage Performance Gap

Robin Harris at StorageMojo links to several presentations from the Seattle Conference on Scalability hosted by Google. Barry Brumitt leads one talk on Using Map-Reduce with Large Geographic Datasets which looked interesting to me.

Best part was that right at the beginning of the presentation he talks about the need for index files to complement raw data, and then...

"Once we have index files, then we load those up onto servers, usually in RAM so that we can actually answer queries from users. It’s really not that useful to have things that are used even slightly frequently on disk because you are limited by the bandwidth to the disk and the number of seeks you can do. So you can do 100 maybe 1000 queries per second on a disk. If you are Google and answering the kind of queries we do, that just doesn’t do it."

Last year we noted another mention of Google's Storage Performance Gap by Luiz Barroso when he outlined the problem at the Intel Developers Conference.

Of course, only some companies have to deal with that scale of queries. And only a select few have the ability to architect a compute infrastructure from the ground up like Google.

But many companies do suffer from I/O bottlenecks and most have longstanding infrastructure in place that cannot necessarily be discarded for a clean slate approach. In those cases, adding speed and performance can be more easily accomplished by using a scalable caching appliance and implementing a centralized storage caching solution from Gear6.

On Bandwidth and Latency

Anil Gupta, author of the Network Storage blog, has written a short series of posts on bandwidth and latency.

Gear6 trailblazing Network Caching

Bountiful Bandwidth Lagging Latency

Where do you focus, Bandwidth or Latency?

Anil has done a nice job framing some of the key issues. Enjoy the reading.

A Closer Look at Databases on NAS

There is a false industry stereotype that NAS can't handle databases well. Many years ago, NetApp and Oracle set out to dispel this myth, supported by a loyal group determined to have the simplicity and ease of use of network attached storage plus their database performance.

Part of these efforts include Oracle and NetApp showcasing the Oracle Austin Data Center as the largest Dell/Linux on earth and the largest NetApp installation on earth with 2.5 petabytes (or 2,500 terabytes) of disk storage, at lease according to this article from 2005: Oracle Technology Network Feature: Running on the Grid

Today, we're seeing more interest than ever in increasing database performance and making the storage component more simple to manage. Earlier this year, we wrote about Oracle users and their attempts at Balancing Capacity and Performance, and more recently about our experience at the Independent Oracle Users Group trade show, Back from Oracle Users Group.

Recently a new blog popped up on the subject:

Gulabani's Databases on NetApp Storage Blog

But the resource with by far the most information:

Kevin Closson's Oracle Blog

The official story from NetApp:

10 Reasons to Use NetApp for Oracle

And another interesting NAS/NFS blog:

Eisler's NFS Blog

We'll be adding to this as appropriate. Feel free to submit other sites in the comments.

Update 6/15/07:

A day after I posted this, Kevin Closson added another fun post to his blog...

Manly Men Only Deploy Oracle with Fibre Channel. Oracle Over NFS is Weird.

Winning the Relay Race

Tony Asaro penned a short snippet about caching and Gear6 in his Computerworld blog. The best part of the post is his explanation that caching is a multi-level process. According to Tony, having more speed helps, similar to a relay race where having a fast runner support a fast runner leads to a win.

Caching has been used throughout data centers for a long time, and is likely to remain a key part of end-to-end architectures. Don't expect cache to completely disappear from individual disks, subsystems, or servers anytime soon. But when large amounts of cache are required, particularly for applications that are otherwise severely I/O constrained, it makes sense to centralize that caching resource into a coherent pool that can be shared across any number of servers and any number of storage devices. We're seen this shift before...remember when disks used to sit directly within or directly-attached to individual application servers?

Blog Watch - Gear6 spotted

If you are a new visitor from Anil Gupta's blog Network Storage, welcome to Thoughtput! Hopefully you'll be back.

Anil has a great post titled Storage Vendors to Watch: Gear6, based on his experiences at SNW. Unfortunately, Gear6 couldn't be at SNW in person as we spent time focusing on the Oracle Users Group show at the same time in Vegas. See Back from Oracle Users Group for a peak at the interaction from that show.

Anil came up with the following customer qualification questions that suit our markets well. There are certainly plenty of workloads other than transaction databases that Gear6 addresses, but this is a helpful starting point.

[begin excerpt from Network Storage]

I found three simple questions that can quickly tell whether someone may want to investigate Gear6 product.

1. Are you using transaction databases?
2. Do you use NFS mounts?
3. Do you have performance issues?

Gear6 product seems to be one of those products that require 10 minutes for presentation, 20 minutes for answering follow-up questions, 30 minutes for demo and then the question When do you want a unit for evaluation?

[end excerpt]

Anil's full post is worth reading, see the complete text and images here.

 

Driving Down Drives

We've been tracking customer and industry interest to reduce the number of disk drives in enterprise environments starting with our earlier post, Drive Reduction. That entry detailed how Turner Broadcasting aimed to simplify their environment by shrinking the number of disk spindles.

A few weeks ago, Byte and Switch released a story about NetApp's focus on power savings...with disks mentioned as a key focus:

"We believe that if more focus is put on getting increased work out of fewer disks while enabling widespread use of higher-capacity and lower-power disks, NetApp can help customers dramatically reduce power consumption."
    -Chris Bennett, vice president, Core Systems, NetApp (Byte and Switch, 3/19/07)

At first, it seemed a bit ironic. How does a company that makes its bread and butter selling disks start promoting disk-reduction strategies? The green angle is one approach too irresistible for most vendors to ignore. We spotted this early on and dubbed the effort Greenvertising.

Now, Dave Hitz of NetApp has provided more clarity on the situation of a storage vendor aiming to sell less disk in this post: Does Helping Customers Use Less Disk Hurt NetApp's Business?

Dave's explanations are clear and grounded in strong business sense, so be sure to read them on your own. It is great to see company's allow their executives to openly share these types of thoughts and strategies.

Big Storage Needs Big Strategies

IDC released a new report sponsored by EMC that the world will generate an incomprehensible amount of data over the next few years. The numbers don't really matter because there is very little way for human beings to understand the scale. Here are a few excerpts from the Associated Press story:

• Add it all up and IDC determined that the world generated 161 billion gigabytes -- 161 exabytes -- of digital information last year
• That's like 12 stacks of books that each reach from the Earth to the sun
• Or you might think of it as 3 million times the information in all the books ever written, according to IDC
• You'd need more than 2 billion of the most capacious iPods on the market to get 161 exabytes

It is kind of IDC to try to put this in terms of a human scale, but can anyone actually comprehend the distance from the Earth to the sun? Or 2 billion iPods?

On another front, Om Malik posted 10 Fun Facts About Storage on GigaOM, that cites a recent New York Times story focusing on the personal storage explosion and how hard drive manufacturers are addressing these needs.

Perhaps the interesting takeaway is how to handle the data deluge. Jon Toigo, who always has an engaging take on things, just launched a new column in Network Computing called the Information Strategist. He recognizes that we need to do more than just acknowledge the growing amount of data, we need to look at ways to rethink how we deal with it.

Data Center Makeovers

There's been some recent interesting reading in the I/O Blogosphere that I wanted to point out.

First, is a comprehensive post from StorageMojo on Architecting the Internet Data Center. This is an inspiring collection of posts on how Internet Data Centers (IDCs) are the new benchmark of IT deployment compared to Enterprise Data Centers (EDCs). IDCs are best represented by the massive data center architectures of Google, Yahoo, Amazon, and Microsoft.

In particular, Robin takes a closer look at how IDCs act as "application delivery engines" and that "Disk I/Os need wise use to optimize system performance." I think these two comments in particular sum up where the industry needs to focus (predictable application delivery) and how we need go get there (efficient and leveraged use of disk I/O).

I've already hinted at the changes ahead in what I call the Great Enterprise Arms Race, which, to better complement StorageMojo, might be renamed the Great IDC Arms Race.

And elsewhere in the I/O Blogosphere

Dave Hitz at Network Appliance had an interesting post on Dave's Blog about how The Network is the Backplane. System architectures have always been a balance between CPU, Memory, I/O, and the Network. Cyclically, specific combinations of these resources gain traction and become de facto architectures for some period of time.

But these primary compute functions continually evolve independently and at times performance tends to favor one over the other. When the delta gets large enough, architectures change. Thinking of the network not as just a network traffic conduit, but also as a system backplane, opens up the playing field for creative thinking and new solutions.