Posts by jms222 298 publicly visible posts • joined 4 Apr 2016
I doubt you can have performance AND low cost AND no endurance problems.
A RAM chip can't say "hang on a minute, I need to sort out my fragmentation problem". It must deliver the data at however many clock cycles it is after the command.
So I think this appears as a (memory-mapped) storage device in a DIMM slot rather than an M.2 or PCIe slot and cannot be used as RAM. Obviously unlike normal storage devices can't do interrupts either so lots of polling needed.
I am well aware of kernel modules and written a few but they typically aren't very deep within the kernel don't help at all when the kernel itself has such an unstable ABI. Only last week I had to fiddle with some userland code, in fact the Python 3.5 socket module that broke because of a change in a minor kernel revision.
This just shows what a heap Linux is. It doesn't have a proper modular device driver system. It just grows including support for every nasty and not so nasty processor and chipset that comes out.
It is also why combining parts from Google and Android device manufacturer is so problematic. The kernel and device support cannot be be separated.
Like others I _used_ to use them (and have a six digit customer number) and as a teen used Dad's credit card for orders and visited their shops in Hammersmith, Edgeware (or was it Colindale) and even Brum but these days they sell mostly very expensive Chinese tat. If I want tat I can get it cheaply online.
We have two shops in Cambridge but I only use them for things I forgot to order online that I must have today.
I am biased but _everything_ should be on ethernet including disks and displays so we can get rid of SAS, Fibre Channel, HDMI, DVI, SATA and possibly even PCIe. which are at some level (packets over SERDES) all the same* and just have different connectors to be a pain in the arse.
Whether object as opposed to block storage makes sense is another thing.
* Save the possible analogue bits on DVI
Why oh why oh why is it it even _possible_ to walk out of these large organisations with data in your pocket. The computers should be set up so that the only way is by using a camera one screen at a time.
If people really need to work on data at home it should be done by connection to the main systems and again they only get to see data a screen at a time.
It's not so much about whether salting and hashing takes place (and remember some authentication protocols require passwords to be stored plainly), it's about removing authentication and putting it in a box called an authentication server.
The authentication server is quite separate from your web server and main databases and _only_ does some very defined things which obviously don't include outputting a password or enumerating accounts.
But then you already knew this if that's your business. Or I hope so.
So given that I'm not about to study the source code what's the difference between me say downloading a kernel with ZFS source separately and building it myself (and wasting resources as a result), downloading a pre-built ZFS module or downloading the kernel with it properly built-in.
Much GPL software was written with reference to versions with another license and they may as well have simply copied it. If Linus had never been born we'd still have BSDs and perhaps less trouble in this area.
Sorry but I've worked in automotive and can see no conceivable reason why you'd need something like that in a car.
I still maintain a system written in Turbo Pascal and runs in a DOS box under Windows Me and yes it's connected to the outside world not via a firewall and no this has never been a problem.
Nothing wrong with COBOL as such. It's the stuff round it that makes a stable system.
If you're advocating using something like node that changes every five minutes you really were born yesterday.
The processing power in modern banking systems goes not on transactions, which are simple and well understood, but fraud detection.
As I have said to one particular colleague of mine.
NEVER EVER pass something user-supplied to a shell. EVER. That includes system(). Of course he went ahead thinking he could escape data himself not realising that no you can't escape single quotes inside single quotes.
I would now add NEVER put user-supplied data into an environment variable that might be passed to bash even if it was supposedly fixed as the fix was a nasty hack. In fact don't even have bash.
Preferably don't use a heavyweight OS at all.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
