Posts by Federal

SaaS Savings are partly from reduced features

The software development process generally follows the "the last 10% of the work takes 50% of the development time" rule. That's where the unique requirements involving customizations and complex data integrations take place. They're hard for developers but greatly simplify work processes for end users.

After the switch to SaaS end users are told "our workflow optimization (to get the overall function to fit into the cloud app) don't allow for that feature in this release."

So the burden of extra work, which was eliminated as a cost of a couple months effort for developers and testers falls onto end users. 2 man months of effort saved at the cost of 5 minutes per week for each of 5,000 people.

If it were only one function set in one app it wouldn't matter, but (at least in our case) the migration of multiple functions to cloud SaaS adds about 2 hours per week to each staff member's workload. Times 5,000 for at least a couple years.

Management claims they've cut IT costs through "Workflow Optimization" and moving to the "the cloud", but staff wind up with less time to work productively and being aggravated by being forced to perform tedious workarounds.

But that's considered progress in a buzzword world.

Paid for "independent" certifications have their limits

Like a prison in which the guards are hired, paid by, and can be fired by, the inmates, the quality of the guarding will be limited. From time to time, there will be escapes.

This applies to ISO certifications and FEDRAMP, too, where the certifiers must market their services through sales pitches to the ultimate recipients of the "independent" certification.

Like driving an airplane on the interstate

NVMe is a very high speed, very low latency protocol that is optimized for on-board short distance data transfers. Just like you can use an airplane to drive down the highway, you can use this protocol over longer distances by putting bridge circuitry at each end of the connection compensate for the different physics of communicating over longer distances. But now you've slowed it down and added costs - like clipping the wings and tail off the airplane to make it fit through overpasses. It would go really fast over some stretches of the highway but then have to slow way down (go through bridge chips) in other places.

TCP/IP is a great protocol for long distances and NVMe, QuickPath, and Hypertransport are great protocols for short distances. Fiberchannel and infiniband are great protocols for medium distances. Many attempts have been made over the years to jam sone of these protocols into different roles and they've generally wound up providing similar (or worse) performance at higher cost (like FC over IP).

Maybe this time it will be different.

They need to be sure the ransomware gets ALL your filles

Make sure it's fully attached and accessible as a "local" set of files.

Wouldn't want the encryption daemon to miss any cloud based backups!

In all fairness

My 3 year old Samsung Note II had a patch from Samsung for the stagefright vulnerability - that was pushed out via an OTA update from T-Mobile.

Which is better than some other mfgs. Are these guys running into a carrier issue rather than an mfg. issue?

Then there's the problem with their N0x emissions

Depending on where you recharge it, in the US, it emits on average 250% of the maximum allowable N0x per mile. And it can be 10 times that depending on how you drive and where you recharge it.

Sooner or later they'll get called on it. The VW cheating has everyone else's cheating being looked at.

The fix is obvious and elegant

Include a stagefright exploit in the daily Google Doodle. Patched phones would be unaffected but other phones will be pwned and accessed at root level, without the need for carrier intervention or jailbreaking/rooting the phone. The exploit will check all potentially vulnerable files. When any such file is found, the exploit will download and install a patched replacement, then reboot the phone.

Everyone goes to Google's home page once in a while, so there would be universal implementation of the patches.

:-)

Android phones have all sorts of selling points (and have sold well, because of the that) but often fall short on security. If RIM can address that weakness, and the phone is otherwise comparable to other Android phones, they might have a winner.

Their biggest competition may come from Windows phone, which also claims some enterprise security features over typical Android phones.

Apple seems to operate in its own segment - I don't see iOS users moving to Blackberry, more secure or not.

NoScript on Firefox

Browsing any other way is Russian Roulette. And even with NoScript, some risk remains and common sense is vital.

But it's hugely helpful. The first few hours of whitelisting pages and scripts you trust to use JS are a little annoying, but after that it's barely noticeable. And it forces you to consider what you've actually gotten yourself into by clicking on that link.

It basically makes the web a series of harmless, static pages, instead of a snake pit of interacting exploits.

Too bad - they're the very best of the Tier 2's

We've used Nexsan units in small quantities for years - their reliability has been stunningly good. They don't have all the features of NetApp and EMC, but they somehow are able to predict drive failures during testing and ship only drives that don't fail. We've had units in the past that used some of the seagate drive models that had a half life of about 18 months in everything else that went almost 7 years in a pair of Nexsan 18 drive boxes before the first failure. To this day I don't get how they're able to do it.

One thing cloud storage is NOT, is less expensive

Cloud costs far more than local storage, but there are many reasons it is a good choice, under some circumstances (very granular increments is one).

But compare the cost of very basic Amazon S3 storage to buying your own:

Enterprise quality NAS / SAN rated drives go for about $50/TB and are warranted for 5 years. Put 24 in a chassis with controller and software and compare that cost to even the very least expensive cloud storage. Now consider the bandwidth (yours and the cloud provider's) charges and I/O charges for accessing the data. S3 pricing is here: http://aws.amazon.com/s3/pricing/

Per GB, cloud costs at least 10 times as much as local storage (up to 100 times or more).

Cloud has lots of benefits, but the "it's cheaper" argument is simply not true. And you still need an IT staff to set up connections to it and to manage it.

It appears it's just the one item

It looks like the only function that's not available is the ability to upgrade HALO software and "observe controller failover." Perhaps Nimbus hasn't worked out how to seamlessly failover through an upgrade on an active system. I'm not familiar with Nimbus/HALO, but it's going to be an operation that's got to execute carefully. If Nimbus' configuratin is active/active, you'd have to first force all operations over to one controller, then upgrade the other, then force operations (these are live, active I/O controller sessions) over to the newly updated controller then upgrade the second controller, then go back to splitting operations between the two controllers.

Our NetApps do this OK, but I'll admit to being impressed by it generally working without much drama. Oracle can't support it with many of its database patches, even in a RAC - you have to shut down the database to patch. (admittedly disk I/O is much less complex than database transactions).

Had to happen sooner or later...

From Heinlen's classic, "The Moon is a Harsh Mistress."

www<.>is<.>wayne<.>edu/MNISSANI/RevolutionarysToolkit/TheMoonIsAHarshMistress.pdf

When Mike was installed in Luna, he was pure thinkum, a flexible logic "High Optional, Logical, Multi Evaluating Supervisor, Mark IV, Mod. L" a HOLMES FOUR. He computed ballistics for pilotless freighters and controlled their catapult. This kept him busy less than one percent of time and Luna Authority never believed in idle hands. They kept hooking hardware into him - decision action boxes to let him boss other computers, bank on bank of additional memories, more banks of associational neural nets, another tubful of twelve digit random numbers, a greatly augmented temporary memory. Human brain has around ten to the tenth neurons. By third year Mike had better than one and a half times that number of neuristors. And woke up.

Am not going to argue whether a machine can "really" be alive, "really" be self aware. Is a virus self aware? Nyet. How about oyster? I doubt it. A cat? Almost certainly. A human? Don't know about you, tovarishch, but I am. Somewhere along evolutionary chain from macromolecule to human brain self awareness crept in. Psychologists assert it happens automatically whenever a brain acquires certain very high number of associational paths. Can't see it matters whether paths are protein or platinum.

("Soul?" Does a dog have a soul? How about cockroach?)

A real shame

We've been using Quantum's libraries for a while now. They're at least as reliable as the 2 other vendors we worked with before switching to Quantum and were quicker to get fixed when the inevitable mechanical glitch showed up after a couple years of use. We've got a bunch of their libraries in production.

I wished they'd just raised prices - we certainly would have accepted an increase after the good experience we've had with their products.

So, where to now? HP? Overland? Qualstar?

It's not always that easy

We had a machine that would crash randomly under heavy load. It could go weeks without having a problem. But if you ran memtest overnight it would show an occasional error.

We replaced the memory and it reduced the frequency of the problem, but didn't end it. The new memory was still showing an occasional error (that was noted as sometimes being a fale positive) when running memtest overnight so we replaced it again.

It turned out that the problem was the hard disk. The electronics on the drive, I suppose. Replacing the hard disk eliminated the problem. The only symptom was memory errors - the hard drive had appeared fine. We replaced the drive so we wouldn't have to worry about losing the user's work. The machine has been flawless for years (with the replaced disk drive).

Never saw that problem before, haven't seen it since. It had 4 very experienced IT staff baffled (including me) for weeks.

Literally hundreds of millions of "parts" need to work perfectly for a PC to work. A problem transistor on one subsytem (motherboard chip) can surface as an error on another.(video card). Or hard disk and memory, apparently - that one still has me somewhat baffled.

But this stuff isn't always easy to figure out. We probably spent a couple $k in staff time diagnosing a bad $100 disk.

This stuff isn't always that easy.