Comments for DrunkenData.com

Comment on Signal and Noise by Administrator

Administrator — Thu, 07 Mar 2013 01:55:02 +0000

Thanks for the feedback. Not sure what to make of the RAMSAM comment, unless these guys know what my transaction load is. If I am doing a million transactions a second, versus say 100,000, how soon would it take to violate the max writes of MLC memory of 250K per cell? How much spare capacity do they offer to substitute or swap in for burned out cell groups?

As for VSC, that sounds right, but I still don’t know about the implementation model. I will be investigating this.

Comment on Signal and Noise by gersc10522

gersc10522 — Thu, 07 Mar 2013 01:40:39 +0000

Jon -
Regarding write wear with RAMSAN, at a recent briefing I was informed you could write to the device at 100% of the write rate for 5 years and not wear out the device. Now, if it is 100% writes I guess we could just have a simple circuit to acknowledge the write and never store anything.

Virtual Storage Center (VSC) bundles all SVC software features except compression together with what was TPC Standard Edition, TPC for replication and adds new analytics around storage tiering.

Comment on Me and Clouds 3 by signal_lost

signal_lost — Mon, 04 Mar 2013 09:35:02 +0000

Rhodesr,
The windows is not actually the SAN/NAS OS, its just used for management functions. FLARE/DART are Linux/Unix OS’s used for SAN/NAS functions. VMAX is not running windows at its core (Think the management or datamovers might use windows, but actual core OS is still custom). John is correct that these systems are open systems with relatively little custom hardware.

“Maybe once CPU cores can be allocated and deallocated at will to different workloads”
I’ve heard this same thing from a AIX admin on the plane the other day to Scottsdale and to try not to laugh.

While VMware does support CPU affinity for 1:1 allocations of physical resources (I think a terrible idea, but an argument for a different day) it also supports hot adding of CPU (If the OS allows, which Windows Datacenter does) and more importantly it also supports MHz based reservations in resource pools. I can dynamically give hard or soft reserve any crazy fractions of a CPU separately of how many vCPU’s I’ve presented at it. I’d argue pooling the MHz and then pooling this is a better play than reserving individual cores, but then again there may be some use case here.

Lastly for people who REALLY want LPAR like management on open systems, Hu and the gang over at Hitachi Data Systems will sell you x86 blades that support LPARs. (which you can then layer a hypervisor on, so you can virtualize while you virtualize).

Comment on Me and Clouds 3 by rhodesr

rhodesr — Wed, 20 Feb 2013 17:23:39 +0000

I know that EMC VNX uses Windows for it’s core OS. I once spent a long debugging sessions with EMC support via a WebX session in the Windows OS on a CX4 array. But I have never before heard that VMAX has Windows in it’s core.

Comment on Actifio and DataCore Software: A Tale of Two Solutions by pipik1199

pipik1199 — Thu, 10 Jan 2013 18:36:23 +0000

Where is the follow-up post(s) on this? I am very curious to find out more about Actifio dependence (and IP legacy) from IBM

Comment on Actifio and DataCore Software: A Tale of Two Solutions by Mox55

Mox55 — Thu, 13 Dec 2012 21:43:33 +0000

All you need to know about Actifio’s architecture, is explained in the
13.4MB IBM SVC Redbook implementation guide- what a frankin freak! Why would I add little appliances into my ENTERPRISE to move data? Seriously, go look at some of the diagrams in that Redbook. Plus Extents? Extents? Plus FC based. Its a hairball. Support will be a nightmare. Its primarily IBM hardware, software, with an APPLE derived interface. I agree its a large market space, but not for such an old design. Nice trojan horse attempt IBM!

Comment on SATA and SAS by URL

URL — Wed, 21 Nov 2012 04:16:41 +0000

… [Trackback]…

[...] Read More here: drunkendata.com/?p=260 [...]…

Comment on Wise and Foolish Builders by mark6330

mark6330 — Fri, 02 Nov 2012 11:17:38 +0000

FYI, I use to work for Citi Group/Bank with its data center in good old Rutherford NJ…..

Comment on Wise and Foolish Builders by mark6330

mark6330 — Fri, 02 Nov 2012 11:16:28 +0000

Amen!!!

Comment on Ranting on SDDC by Ernst Lopes Cardozo

Ernst Lopes Cardozo — Thu, 01 Nov 2012 23:24:28 +0000

The “time honoured” top-down design method was abandoned when Mainframes came up: a new business requirement was dealt with by building or buying a new application that was supposed to run on the existing kit. Today, building a new compute infrastructure (storage, backup, servers, networks, etc.) takes many ‘business clock ticks’. If companies have to wait till new hardware has been specified, acquired, installed and tested, the opportunities will long be gone. The design method is bottom up as high as possible without knowing exactly what the business will want and need, then top-down as soon as the business requirement can be formulated. In practice, this means that not only the compute platform (be it mainframe or hypervisor) including storage, BC, DR but also databases, message queuing and all the management instrumentation are in place, running existing processes, with enough spare capacity to accommodate all but the most extravagant new business requests.
I am not fond of such acronyms as SDDC either, but in this case there is some substance behind it. Software will take the place of our IT heroes a.k.a. system administrators, that until now controlled server parks with their bare mouse. IT is finally applying its own recipe to automate its own business.