Posts by dakra

Replace DRAM? Think bigger! Replace SAN!

This will be a major big deal for large systems, especially with a petabyte or exabyte of register addressible memory, some pooled and some shared within a cluster.

CXL will do for aggregated, shared, and pooled large memory what SAN did for storage.

SANlock semantics will provide for locking within shared memory.

CXL Switches will provide intersite memory replication for disaster recovery. That will be analogous to SAN switch-based intersite storage replication.

Programs could access shared files through virtual memory mapping semantics while others use the old I/O syntax, libraries, and system calls.

IBM:

* IBM could extend its lead in clustered shared memory with Parallel Sysplex and Geographically Dispersed Parallel Sysplex for files in memory. Some will imagine this as clustered System/38 Single Level Storage.

_ _ Mainframe Coupling Facilities, VSAM, DB2, spooling, MQ, DiV, GRS, GDPS etc. could be enhanced to exploit and support CXL persistent memory.

* IBM could ignore the possibilities and simply let the rest of the industry catch up to and then surpass what mainframe clustering can do.

Why bother testing? What does it do if the test fails?

Reaching a popular service does not prove that the device can reach the services it might actually need to reach.

Why doesn't the device only test to see if it can reach the services it might need to reach?

Even reaching the services it might need someday does not prove that it will again be able to do so when it needs to.

-

What does the device currently do if the reachability test fails?

What does the device do if it can't reach what it needs to when it needs to?

Don't blame Intel. Blame the customers who didn't value valid results.

Intel didn't kill ECC and parity memory in consumer and business end-user PC's. Rather the customers and clone makers did, and the trade press was complicit.

Early personal computers did not have either parity or ECC memory. Then the IBM Personal Computer came out for business use, with advertising referring to it as "The IBM of Personal Computers." IBM's middle name is Business, as in "International Business Machines."

IBM PC's carefully tested all memory at Power On. That took time which users did not appreciate. The PC stopped hard if it encountered any, even momentary error while running. People did not appreciate losing their unsaved work in progress.

Clone makers won business by doing several things:

* They offered BIOS settings to skip the memory test at power-on. People loved the time savings.

* They offered computers without parity memory at a lower price. People loved the lower price.

The trade press was complicit. It made fusses about all sorts of things, but their editorials did not educate users about the risks of skipping tests and not having parity checking. Furthermore, their reviews of these clones did not downgrade them for lacking parity memory or offering an option to skip the power on test.

During those years, IBM strategically shifted from telling customers what is good for them to being "Market Driven." That meant giving customers what they want. Even business and health care customers voted with their pocketbook that they did not value valid results. The same IBM executives who said it was a bad decision technically, said it was the right market driven business decision to drop parity memory from desktops and laptops. I heard this directly from executives of both the PC and the memory chip divisions sitting together addressing an internal IBM audience.

Sysadmin sank IBM Mainframe by not going one VM deep enough

The article's headline got it wrong.

Don't say the "Sysadmin sank IBM mainframe by going one VM too deep"

The problem isn't that he went one VM too deep. It's that he went one VM too shallow. He killed the first, top level VM.

If he killed the second level VM, it would have taken its children along with it, but left the top level alone.

The headline could read, "Sysadmin sank IBM Mainframe by not going one VM deep enough."

IBM 3800 maxed out at 12,423 lines per minute, not 20,000

The 3800 printed at 20 inches per second. It could print pages sideways, where the paper was landscape mode, and so was the type, so it looked like portrait mode, after the continuous paper was burst into separate sheets, and trimmed of its sprocket holes into 8.5 by 11 inch sheets. That 12,423 line per minute number comes from

(11 inches of print lines per page )

* (8 lines per inch)

* (60 seconds per minute)

* (20 inches per second)

/ (8.5 inches per page)

= 12423.5 lines per minute.

A major reduction on throughput was how long it took the operator to change a box of paper. It could print a box in about 15 minutes, but took another 5 to replace it. The solution was to use a non-IBM paper handler. It held a very large roll of paper outside the printer, with rollers to pass the paper into the printer and twist it around to feed correctly.

""If you actually want a STEM job, be an engineer: science graduates are mostly under-employed..."

A classic explanation of the difference between scientists and engineers says, "Scientists make it known, engineers make it useful."

Most employers want to make money via things that are useful.

Most software development is for useful business applications with practical heuristics or operational scripts. They mostly exploit, but do not develop office suites and graphics package, compilers, interpreters, other end user generic applications, and infrastructure middleware (server daemons of all kinds). Fewer still do computer science, generating academic proofs of theorems of orthonormality, completeness or complexity. I was successful in the industry for over three decades before needing to learn and use "regular expressions," regarding which the Computer Science Graduate Record Exam devoted several questions.

For my undergraduate education, I chose a university where I could switch from Physics to Engineering without a problem. I didn't wait until my sophomore year. Programming was something I did as needed for a business, charitable, or personal purpose.

Don't major in Computer Science, unless you want to be an academic. Major in a domain that is interesting to you, that you will consider fun and fulfilling to work in, and will pay the bills. Learn enough math, propositional logic, queuing and graph theory, and lastly programming, to further your real interest.

"The supported lifetime of the device" is already over, but I am upgrading anyway.

According to Microsoft and HP, my quad core i7 8GB full HD Nvidia+1GB laptop is already out of support for WindoX. The computer is from before 2013 and so was not tested for compatibility. (It is from 2009) New equipment is recommended. However, when I ran the preupgrade compatibility check, it only complained about an old version of a partition manager program, which I uninstalled. Device Manager for both Windoze7 and WindoX show an error flag only for the IntelTurboBoost Driver still dated from 2009.

Under WindoX, everything else works, even PDF Redirect, MS Office XP, and MSW Logo from 2002, except for my old DOS 16 bit applications, which no Windows x64 version can run on its own, I was very surprised that it did not overlay the Grub2 bootmanager and its settings for linux/Windows dual boot.

WindoX takes 37% more memory than Windoze7 doing nothing once stabilized after boot (2.6 vs 1.9 GB). If I had less memory, that would be a show stopper.

I will keep WindoX for some of the reasons given in previous comments, and also because (1) It supports my Bluetooth headset as both stereo headphones and as a phone headset with its various control buttons. and (2) It boots faster, only 50% longer than linux to the login prompt. I really only run Windows a few times a year, but I take quicker boot as an indication of more efficient performance.

Distributed system ignoring error messages from its partner

Two custom programs talking to each other across a network.

Sometimes, transactions from the client were not being recorded at the server online. There was a batch process to upload all the day's transactions at end of day, so they all eventually got there. This problem of occasionally "not getting" online transactions went on intermittently for over a year.

I looked at the server side code, and found that problematic transactions would be sent to a named destination. I asked the programmer where that was. He didn't know, it was just a name. I asked the system administrator, who told me it was the console log. I asked where that went. He told me, to check the job control language. That indicated it went to print. I asked the clerk where the printouts went. She pointed to a cabinet. I looked at the previous week's printouts. Did you lose a transaction from this account on this date? and that account on that date? Yes, how did I know? Uh, it's right here on the printout.

I went back to the server code and saw that if there was a problem it not only logged the request, but also sent back an error code to the client.

I looked at the client side code.

On the client side, the coding technique to catch error indications sent from the server was to register the address of the routine that would handle the error. Unfortunately, the client developer had never implemented any error handling routine.

At the same time, I pointed out to the programmers how some of the code would fail in Y2K. They said they had no intention of working there for the next 22 years.