Solid state disks (SSD) are data storage devices that use nonvolatile memory (Flash) and volatile memory (SDRAM) to store data.  SSDs are gaining popularity and replacing hard disk drive (HDD) in ultra mobile PCs (UMPCs) and notebook PCs because SSDs have no moving parts, making them rugged, immune to vibration, and have no altitude operating restrictions.

SSD controllers treat Flash memory as a block device that allows fixed-size data blocks to be read and written, much like disk sectors. This allows standard file systems designed for magnetic disks, such as FAT (File Associate Table), to utilize Flash devices.

Flash: longevity issues
Mapping the blocks of data onto Flash addresses in a simple linear fashion presents two problems. First, when the file system in mapped onto a Flash device, frequently used erase units wear out quickly, slowing down access times, and eventually burning out. This problem can be addressed by using a more sophisticated block-to-Flash mapping scheme and by moving around blocks, referred to as wear-leveling.

The second problem is writing data blocks smaller than a Flash erase unit. Suppose that the data blocks that the file system uses are 4 KB each, and that Flash erase units are 128 KB each. If 4 KB blocks are mapped to Flash addresses using the identity mapping, writing a 4 KB block requires copying a 128 KB Flash erase unit (block) to RAM, overwriting the appropriate 4 KB region, erasing the Flash erase unit, and rewriting it from RAM. Furthermore, if power is lost before the entire Flash erase unit is rewritten to the device, 128 KB of data are lost. Wear-leveling can also address this issue, but adds complexity and cost to the design process.

F-RAM: high endurance, no leveling needed
In SSDs, the endurance and wear-leveling required in Flash virtual-block-to-sector map (hot data) can be eliminated altogether by storing the virtual-block-to-sector map in a F-RAM device.

F-RAM has virtually unlimited endurance (1E+14) and is writable at byte granularity and there are no block-erasure constraints like NAND Flash.

Contact an F-RAM applications engineer to find out how F-RAM can improve your next solid state disk design.

Email us at framinfo@ramtron.com or call 719-481-7000.

Network security appliances are installed in-line on a network, thereby all network traffic must traverse it. If the security appliance were to fail, network connectivity could be lost. Network security appliance vendors have solved this issue by implementing an option to “fail open” or “fail closed.” This is a controlled setting, by which the user can decide whether or not network traffic should pass in the event that the security appliance fails.

If the system fails, the user can configure the system to fail open (no traffic passes through the system) or fail closed (all traffic passes through the system). Thus the user has complete control over system failures. In designing network security appliances, fail open (LAN bypass) refers to a system which, when it suffers a critical failure or overload, drops to a lower level of security, allowing traffic to pass without checking it rigorously.

LAN bypass function provides network stability and prevents network problems when platform shuts down. The following are potential failure events that can trigger LAN bypass:

1. Device hardware failure
2. Power failure
3. Internal application failure
4. Operating system crash

There are two communications states for the LAN bypass function: Normal State and Bypass State. A watchdog timer (WDT) is used to control and switch the communication between the two states.

Why use F-RAM State Saver?
The LAN bypass function is used to link (or short) two independent Ethernet ports when user’s application software halts or when the power is switched off. In a conventional implementation, the command to the relay driver is memorized every time its state has been changed (D-flip flop + a lithium button cell battery such as CR2032) making sure the LAN bypass function stays at its original state after the appliance is power cycled.

Nonvolatile F-RAM state savers eliminate the use of backup battery in a “LAN bypass” circuit, ensuring no network disruption on power up. F-RAM state savers also save board space, as compared to a latching relay that is bulky and expensive.

Contact an F-RAM applications engineer to find out how F-RAM can improve your next network appliance design.

Email us at framinfo@ramtron.com or call 719-481-7000.

A router is a networking device whose function is to forward digital data packets to the appropriate parts of a computer network. Routers are found at every level in a networked envrionment. For example, routers in access networks allow homes and small businesses to connect to an Internet Service Provider (ISP). Routers in enterprise networks may link thousands of computers within a campus or enterprise. On a larger scale, routers in the backbone are not usually directly accessible to end-systems. Instead, they link together ISPs and enterprise networks with long distance trunks.

Understanding router memory
At the hardware level, there are five main types of router memory: DRAM, EPROM, BBSRAM, Flash, and EEPROM. Each memory type serves a specific data storage function. For example EPROM is used for as a boot ROM to load crucial firmware components. DRAM is typically used as processor memory for routing tables, protocols, and network accounting applications. Flash memory stores Operating Systems and multiple system software.
BBSRAM is commonly used in routers for permanent storage of the system configuration scripts, patches, and logs that are writeable. This implements any configuration changes saved since the router module was last reset. BBSRAM is also used for permanent storage of hardware revision and identification information, as well as Media Access Control (MAC) addresses for LAN interfaces.

When routers are switched on, they copy the Operating System from Flash memory and copy the complete configuration setups from BBSRAM, which are copied and stored in the DRAM until the router is turned off. Routers then configure the interfaces and begin building their routing tables. Forwarding of datagrams commences when the router has enough information to determine where to send the data packets.

Why use F-RAM?
The lithium battery poses the weakest link in BBSRAM. Battery field failure due to low charge, corrosion, or vibration can cause data loss in the memory, threating a system failure in the router.

Nonvolatile Ferroelectric RAM (F-RAM) offers an ideal and reliable alternative: it requires no external components such as battery or capacitor and provides instant, reliable data saving.
During power failure, data is preserved in the BBSRAM by using V^_BAK — as long as the battery power supply is working.

In F-RAM, data is stored natively in the nonvolatile ferroelectric cell. There is no need for a battery and the stored configuration or log settings can be read immediately from the memory upon power restoration.

Contact an F-RAM applications engineer to find out how F-RAM can improve your next network appliance design.

Email us at framinfo@ramtron.com or call 719-481-7000.