« February 2010 | Main | April 2010 »
Posted at 02:35 PM | Permalink | Comments (3) | TrackBack (0)
Reblog (0) | | Digg This | Save to del.icio.us |
For years my typical knock against all the people inside EMC running around thinking I should bow down to the magnificence that is VMware was “So I’m supposed to get excited about selling shipping containers for operating systems?”
They didn’t like that and they didn’t like it because I was right.
A virtual machine is a shipping container for whatever much more interesting thing was placed in it.
And that was the case for a long while until VMware started changing.
Taking my same Virtual Appliance I used to have a look around Redis I downloaded the 90 Day evaluation of tc Server 6.0 from SpringSource and took a look at that. No wget this time since it’s behind a thankfully brief registration form with the full tc Server bundle being 170MB in size. A hell of a lot smaller than WebLogic or WebSphere but orders of magnitude larger than Redis.
After running install.sh I was presented with a bunch of options all appearing to offer me the ability to install a host of stuff I probably wouldn’t need, so after just selecting the Application Server I hit a bump when it wouldn’t install into a directory which wasn’t already created. mkdir solved that problem and then we were away.
tc Server has a hell of a lot more options than I had expected and it took me a bit to figure out what was where, this is what I get for relying solely on the kindness of the –h flag instead of reading the docs but I created two server instances, checked they were listening and then modified their server.xml files to configure a session replication cluster.
I could have probably created 200 or 2000 instances and done the same thing but two struck me as enough and it saved on keystrokes.
Now on my colourful Guest VM I had a clustered application server and a high performance in memory database running at the same time. Not being a developer I didn’t have an app to take advantage of any of that but probably the longest part of the entire process was downloading the tc Server installation files.
"Well bravo Zilla you made a bunch of workload proven code do what it should do out of the box, so therefore you're now an expert on everything." Hardly, but the point is all this stuff is now part of what’s becoming a VMware Suite and I'd have no cause to look at it were that not the case.
I think that's going to be true of a lot of people. I think a lot more eyeballs are going to focus on these now.
Lets look at the new suite of VMware technologies from top to bottom, had I artistic ability I’d draw it but I don’t so you can visualise it with nice shapes and colours in your own head as you read it off this chart.
Now doesn't that look a lot more interesting than running the contents of a VMDK file and moving it around the place? So, how far away are the rest of us from provisioning an application in vCenter and watching as tens, hundreds or thousands of configured instances come online to support the application workload?
Not that far.
The hyperscale people already operate in some similar fashion but that way of running your infrastructure is on the cusp of coming down to the folks with more modest needs and who don't have an army of people with PhDs working in their IT department.
I think technologically there's a gap to be filled with the concept of the Guest OS but I'm not Paul Maritz and therefore being mortal I could be wrong. Right now you bring your own OS, be that Linux, Windows, Solaris, BSD or so on and while I'm not advocating VMware (or EMC) going shopping for a distro vendor I think we'll see some form of side approach to a VMware specific Guest OS the way we've seen it in other places on that chart.
Were it to occur all I can guess is that it'll probably be Open Source.
Sure you can still bring your own but if you don't care here's Just Enough of our OS to run all the bits above it on the chart and we'll take care of the licensing, deployment, support and patching.
We'll see one way or another I suppose. If it's actually a gap it'll get filled soon.
Posted at 02:41 PM | Permalink | Comments (0) | TrackBack (0)
Reblog (0) | | Digg This | Save to del.icio.us |
Hiring the Veritas talent on the open market.
Reqs open in Engineering for a number of positions.
(Some in file system development)
Posted at 11:01 PM | Permalink | Comments (0) | TrackBack (0)
Reblog (0) | | Digg This | Save to del.icio.us |
I had wondered aloud more than once if VMware would feel compelled to buy a database for their growing middleware business what with Microsoft, IBM and Oracle all having their own. If they did I imagined it would be an Open Source offering as that’s pretty much how the company is evolving.
Maybe they’d work on Postgres? Maybe they’d buy EnterpriseDB? What I didn’t expect is that they’d leap over all of that and start commercially funding Redis.
What I know about databases I know just from getting my hands dirty and having a go when the opportunity arose, but to me Redis doesn’t look like SQL Server or DB2 or Oracle or MySQL.
For one thing after I started up my Mint 8 Virtual Appliance, which I’d grabbed off VMware’s Virtual Market Place, I downloaded the latest stable release of Redis from Google Code using wget. We’ve all seen DB installs so I expected this would take a while, a little over 2MBs of tarball later it was sitting in my home folder.
I did wonder if I had grabbed the wrong file. Maybe I downloaded just the documentation? Nope. Correct file.
Unzip and untar and you end up with a bunch of files in a directory. There’s no autoconfig or installation script so after setting make loose on them out popped a couple of binaries. I started up the Redis server and watched it listen on port 6379, connecting to that port via telnet I could send it commands and get responses but I spotted Redis comes with a CLI binary which allows you to do the same thing.
Using set and get commands I found I could store and retrieve values. okay so right now it sounds like someone’s first attempt at a database, it stores and retrieves data. Woo! But here’s the thing it appears to be blazingly fast as it’s written in C and it’s an in-memory database.
Everything sits in RAM with writes being written out in asynchronous bursts or in an append only format.
It also supports replication. It’s not Oracle DataGuard but you can’t set up DataGuard by modifying one line in a configuration file!
So, are you going to run SAP on this? Right now perhaps not. What about your distributed computing cloudy app? Oh yes and there’s Java support so you can see a touch point for the Spring Folks.
I think I see what VMware is doing here, they’re jumping over the traditional apps requiring the large DBs on the big tin and going after the hyperscale speed freaks. Twitter and the like could probably be implemented in Redis. Keep the DBs in memory and scale out across numerous servers all while keeping up with the voluminous store and read operations a site like that has.
I like this. I think it’s more important in the long term than a lot of us have grasped.
How all of this comes together, Redis, Spring, Zimbra, and how it’s delivered with VMware as a suite will be interesting to watch.
Posted at 02:35 PM | Permalink | Comments (2) | TrackBack (0)
Reblog (0) | | Digg This | Save to del.icio.us |
It strikes me that the role of the CMO is a tough job. Marketing typically takes the hit faster than product development or sales and yet it can’t not take a hit if either product development or sales aren’t executing the way they should be.
So you learn to do what you can when you can.
Typically EMC marketing has operated like the Nac Mac Feegle clan from a Terry Pratchett novel. Attack every target, headbutt anyone who isn’t a partner and take every hill all at the same time. You’ll take your lumps in the process but if you succeed more than you fail you’ve done the job.
This however leads to some contradictions and things which don’t quite join up. Every EMC employee can explain what it is the company does, but you’ll find as many variations as there are divisions. How does this fit with that? Unless you’re an expert on this and on that you can end up drawing a blank.
Not everyone are those folks in the company who can buffer many if not all of the moving parts in their head. A lot of people are just focused on doing their job to the best of their ability and getting deep into the other stuff detracts from that.
So it’s with interest I read that Jeremy Burton has been pointed Chief Marketing Officer. I know this much >-< about Jeremy outside of the fact he’s a fan of social media but feedback from people outside of the company appears to be quite positive.
This is encouraging when I look at some of the messaging challenges the company has now and what I’d imagine will be required over the next two years.
GERONIMO!
Posted at 03:31 PM | Permalink | Comments (2) | TrackBack (0)
Reblog (0) | | Digg This | Save to del.icio.us |
Simple math thus says that roughly 12.5% of USP-V and –VM controllers are actually virtualizing 3rd party storage (25% of 50% = 12.5%).
(By the way, if we took Hu literally, we could assume that the 50% that are NOT "virtualization enabled" are in fact AMS' and the rest are USP-V's, thus making the percentage of USP-V and VMs that front 3rd party storage a mere 6.25%.)
via thestorageanarchist.typepad.com
This isn't me taking a pop at Hitachi I'm just asking a question. If you're running a USP/XP are you using Universal Volume Manager or External Storage XP? If not feel free to say so, if you are feel free to say what you're using it to do?
I've run into more than my fair share of accounts where they're running a USP, their account rep has bombarded them with the licenses required to virtualise external storage but it never happens. Or they start, do a bit and leave it at that.
Anyway, just a question. If you're not you're not, if you are how are you using it?
Posted at 02:25 PM | Permalink | Comments (4) | TrackBack (0)
Reblog (0) | | Digg This | Save to del.icio.us |
Pat Gelsinger on EMC Virtual Storage.
Here’s an uncommon point, some of you have seen Global Federation in action already.
More later.
Posted at 08:37 PM | Permalink | Comments (0) | TrackBack (0)
Reblog (0) | | Digg This | Save to del.icio.us |
First thing I have to say is that when I first installed Win7 on my old workstation, I learns is that powerpath does NOT work correctly at all on Windows 7.
Second thing I learn is that Windows2008-R2 64Bit almost perfectly emulates Windows7 when you put it in desktop mode and enable all the bells/whistles.
Third thing I learn is that 8CPU cores 16G’s of ram, and 2 1Gbyte video cards makes World of Warcraft SCREAM.
Even when there are four VM’s running in the background.
Yes, I’m *THAT* nerd.
![]()
But the best part of it was the migration. Now as I’ve said in the past I’ve been running the boot-from-san for some time (most recently with the Win2k8R2/Powerpath up and running), so of course the Emulex drivers were already a part of the operating system. So this is how it went:
1. Build new system.
2. Shut down old system.
3. Move Video/Emulex cards to new system
4. Connect and Power On.
5. Reboot twice as motherboard/CPU specific drivers are loaded.
6. Done.
Total migration time – about 45 minutes, including hardware swap.
Now *THIS* is the reason I strongly support and encourage boot-from-san in a datacenter. Not only does it make it amazingly easy to protect your data. (SnapView, MirrorView, etc) but you have the option of upgrading hardware and keeping your disks/OS in-tact.
I suddenly have the urge to build a system with 16GB of RAM and to lobby the PowerPath team to support Windows 7.
They probably won't as it's a desktop OS and they're busy on something else right now but hope springs eternal.
Posted at 01:30 AM | Permalink | Comments (0) | TrackBack (0)
Reblog (0) | | Digg This | Save to del.icio.us |
One of the things which interests me about the industry move to reverticalisation, a return to the Austin Powers style of 60’s computing where everything in the data centre comes from one supplier, is that the focus changes from servers, storage and networking back to applications using all of that.
You’ll have to remember concepts like zoning, masking and RAID groups are pretty wonky things to IT Generalists or Application Administrators. Indeed they’re pretty wonky no matter how you look at them.
I walk around the place and there’s this huge disparity in who has what skills in their IT shop and yet to deploy applications on a shared storage architecture it’s expected you’ll be up to speed on not only the Application itself but also the storage array it’ll reside on and the interconnect you’ll attach them with.
We expect everyone to be at a baseline across multiple technologies when for the most part the people involved are looking for this much capacity to support this workload and whatever voodoo has to go on to make that happen should be something they don’t have to think about.
Where’s the value add for the Applications person also having to be the Storage person?
Posted at 01:24 PM | Permalink | Comments (0) | TrackBack (0)
Reblog (0) | | Digg This | Save to del.icio.us |
This "trusted computing infrastructure" concept will be based on a hardware root of trust that will come with Intel's latest chip and encryption technology, secure virtualization environment, security information and event management, and GRC (governance, risk and compliance) management software, the companies said in a joint statement. This comes after EMC's January acquisition of Archer Technologies, which provides GRC software.
EMC acquired RSA in June 2006 and VMWare in December 2003.
"Today, most organizations have little to no visibility of what's occurring within the infrastructure layers of clouds, making it impossible to verify their security," Pat Gelsinger, president and COO for EMC's information infrastructure products, said in the statement. "Together, our companies are demonstrating that internal and external clouds can be visible, measurable and reportable for the secure management of a company's most important business processes."
I'll have to think deeper about this when I get a moment but I can see where you could secure a Vblock from the silicon upwards.
Posted at 03:09 PM | Permalink | Comments (0) | TrackBack (0)
Reblog (0) | | Digg This | Save to del.icio.us |
