3 ms·

I wish SD cards would implement a key-value storage interface natively. It would instantly remove the need to implement a filesystem in many embedded systems eg

by mrb 13y ago

I wish SD cards would implement a key-value storage interface natively. It would instantly remove the need to implement a filesystem in many embedded systems eg. music players: all they need is access to keys (song filenames) and values (blob of ogg/mp3 data).

kamaal 13y ago
As some doing a bit of embedded system work these days, I was wondering why the MCU manufacturers don't offer a key value store(Even small ones would do) for configuration purposes. The most famous ways of managing configuration is serializing a structure on EEPROM/Flash, or writing a string with lengths of the strings as delimiters. Even if you assume, its for saving space etc. The way I see you will inevitably use space while you write the code to serialize and de serialize the configuration data.
- michaelt 13y ago
  Well, a lot of microchips tell you to write a value to a specific location in memory because that location in memory is physically wired to the hardware you're controlling. So for example the output pins of a microcontroller are a memory location wired (through a buffer and some control circuitry) to the output pins. The PWM circuit is a counter and a comparator, where the comparator's inputs are the counter and a memory location. You could write a key/value to memory location mapping layer, of course, but that's basically what vendor libraries, device drivers and operating systems already provide.
saljam 13y ago
These hardware KV interfaces are still fairly low-level. Fixed-size values, etc. So you'll still need to have some abstraction layer that handles that. You might not call it a filesystem but I bet it'll look a lot like one.
rektide 13y ago
SD cards implement a linear map of blocks. We typically introduce a hierarchical key-value store on top of that- a filing system where we file values by certain keys. What is your ask? Get rid of hierarchy and use only a single flat directory on the SD card? Plan9 was close to the kind of vision you describe- configuration and state for applications lived live in the file system.
zwieback 13y ago
That would be nice and it's doable. I like eMMC for microcontroller apps since it implements wear leveling and bad block management internally. Adding a simple key-value mapping could probably be added without too much effort. eMMC isn't meant to be removable, though.
Someone 13y ago
Classical Forth systems worked that way. Being Forth, they of course went for minimalism here. Keys were unsigned ints (actually, ints interpreted as unsigned) and all values were 1024 bytes (see http://c2.com/cgi/wiki?ForthBlocks http://c2.com/cgi/wiki?ForthBlocks) Programming that way was fun, but I wouldn't want to use it on a system with megabytes of RAM. Embedded, it would be fun to implement what you describe on top of that, though.
TheLoneWolfling 13y ago
Database as a file system. In some ways it actually makes an odd sort of sense... SELECT * FROM sdc WHERE Type='mp3'; I could see uses for something like that. You could even treat it like a traditional file system for fallback purposes, if one of the tags was a 'directory' tag. Also, it would make sense in cases where you have... [whatever the equivalent for NUMA is for disks. NUDA? Things like hard drives with a limited flash cache.] Store the indexes on the flash or in RAM (periodically backed up to the disk, of course). Biggest issue would be wear on the flash, though.

by mrb 13y ago

kamaal 13y ago

As some doing a bit of embedded system work these days, I was wondering why the MCU manufacturers don't offer a key value store(Even small ones would do) for configuration purposes. The most famous ways of managing configuration is serializing a structure on EEPROM/Flash, or writing a string with lengths of the strings as delimiters. Even if you assume, its for saving space etc. The way I see you will inevitably use space while you write the code to serialize and de serialize the configuration data.

michaelt 13y ago
Well, a lot of microchips tell you to write a value to a specific location in memory because that location in memory is physically wired to the hardware you're controlling. So for example the output pins of a microcontroller are a memory location wired (through a buffer and some control circuitry) to the output pins. The PWM circuit is a counter and a comparator, where the comparator's inputs are the counter and a memory location. You could write a key/value to memory location mapping layer, of course, but that's basically what vendor libraries, device drivers and operating systems already provide.

saljam 13y ago

These hardware KV interfaces are still fairly low-level. Fixed-size values, etc. So you'll still need to have some abstraction layer that handles that. You might not call it a filesystem but I bet it'll look a lot like one.

rektide 13y ago

SD cards implement a linear map of blocks. We typically introduce a hierarchical key-value store on top of that- a filing system where we file values by certain keys. What is your ask? Get rid of hierarchy and use only a single flat directory on the SD card? Plan9 was close to the kind of vision you describe- configuration and state for applications lived live in the file system.

zwieback 13y ago

That would be nice and it's doable. I like eMMC for microcontroller apps since it implements wear leveling and bad block management internally. Adding a simple key-value mapping could probably be added without too much effort. eMMC isn't meant to be removable, though.

Someone 13y ago

Classical Forth systems worked that way. Being Forth, they of course went for minimalism here. Keys were unsigned ints (actually, ints interpreted as unsigned) and all values were 1024 bytes (see http://c2.com/cgi/wiki?ForthBlocks http://c2.com/cgi/wiki?ForthBlocks) Programming that way was fun, but I wouldn't want to use it on a system with megabytes of RAM. Embedded, it would be fun to implement what you describe on top of that, though.

TheLoneWolfling 13y ago

Database as a file system. In some ways it actually makes an odd sort of sense... SELECT * FROM sdc WHERE Type='mp3'; I could see uses for something like that. You could even treat it like a traditional file system for fallback purposes, if one of the tags was a 'directory' tag. Also, it would make sense in cases where you have... [whatever the equivalent for NUMA is for disks. NUDA? Things like hard drives with a limited flash cache.] Store the indexes on the flash or in RAM (periodically backed up to the disk, of course). Biggest issue would be wear on the flash, though.