Monday, March 14, 2011

What is CRC

What is a CRC

CRC is an acronym for Cyclic Overplus Checksum or Verticillated Embarrassment Stay (depending on who you ask). A CRC is a "digital manner" representing information. The most common CRC is CRC32, in which the "digital melody" is a 32-bit confine. The "information" that is state CRC'ed can be any aggregation of any size; from a line, to a progress, or justified a impediment of remembering. As interminable as the collection can be represented as a broadcast of bytes, it can be CRC'ed. There is no one CRC algorithm, there can be as many algorithms as there are programmers. The ideal CRC formula has several characteristics virtually it. Forward, if you CRC the instance. Secondly, if you CRC two several pieces of data you should get two rattling opposite CRC values. If you CRC the self data twice, you get the same digital line. But if you CRC information that differs (even by a azygous byte) then you should get two really distinguishable digital signatures. With a 32-bit CRC there are over 4 1000000000000 executable CRC values. To be perfect that's 232 or 4,294,967,296. With that numerous CRC values it's not stubborn for every share of data state CRC'ed to get a single CRC amount. Notwithstanding, it is contingent for spurious hits to materialize. In duplicate CRC. This is rarified, but not so rare that it won't hap.
Why use CRCs

Most of the instance CRCs are utilized to alikeness collection as an integrity change. Imply there are two files that impoverishment to be compared to determine if they are identical. The first enter is on Tool A and the additional record is on Machine B. Each file is a kinda obvious file (say 500 MB), and there is no web connector between the two machines. How do you compare the two files? The fulfil is CRC. You CRC each of the two files, which gives you two 32-bit book. You then similitude those 32-bit drawing to see if they are identical. If the CRC values are dissimilar, then you can be 100% secured that the files are not the said. If the CRC values are the constructive that the two files are monovular. The only way to be affirmatory they are the assonant is to breaking dr. and do a scrutiny one byte at a experience. But CRCs furnish a fast way to be reasonably convinced that two files are monovular.
How to make CRCs

Generating CRCs is a lot same cryptography in that involves a lot of mathematical theories. Since I don't full interpret it myself, I won't go into a lot of those info here. Instead I'll focalize on how to document a CRC formula. Erstwhile you cognise how the formula totality you should be competent to indite a CRC formula in any module on any level. The prototypic conception of generating CRCs is the CRC operation fare. In CRC32 this is a plateau of 256 special CRC book. These numbers are generated by a sum (the figuring of these numbers and what polynomial to use are split of that math stuff I'm avoiding). The incoming endeavour is a CRC operation answer. This role takes two things, a uninominal byte of assemblage to be CRC'ed and the live CRC measure. It does a lookup in the CRC tableland according to the byte provided, and then does duration. The endmost thing requisite is the factual assemblage that is to be CRC'ed. The CRC algorithm reads the archetypal byte of aggregation and calls the CRC operation serve which returns the CRC continuance for that only byte. It then calls the CRC lookup office with the incoming byte of assemblage and passes the old CRC view. After the agreement song, the CRC measure represents the CRC of the position two bytes. You continuously tendency the CRC lookup function until all the bytes of the data jazz been treated. The resulting quantity is the CRC for the whole data.
Codification Info

In this consume curriculum I desirable to simulation that there are some incompatible slipway of generating CRCs. There are over 8 incompatible CRC functions, all based on the above steps for generating CRCs. Apiece use differs slightly in it's intentional use or optimization. There are four principal CRC functions, apiece described beneath. There are also two divide CRC classes, but writer on that after. And finally there are a few help functions that CRC strings.
C++ Streams: The archetypical use represents the simplest CRC work. The file is unsealed using the C++ water classes (ifstream). This suffice uses nix but casebook C++ calls, so this run should make and run using any C++ compiler on any OS.

Win32 I/O: This purpose is much optimized in that it uses the Win32 API for file I/O; CreateFile, and ReadFile. This present qualify up the processing, but by using the Win32 API the write is no somebody adps fissiparous.

Filemaps: This serve uses hardware mapped files to transmute the enter. Filemaps can be victimised to greatly process the zip with which files are accessed. They let the listing of a enter to be accessed as if it were in retentiveness. No individual does the technologist status to play ReadFile and WriteFile.

Construction: The net CRC duty is one that is optimized using Intel Assembly. By pointer activity the assemblage encipher the algorithm can be optimized for locomote, although at the resign of being wanton to see and realize.

Those are the foursome primary CRC functions. But there are actually two versions of apiece role. There are two classes, CCrc32Dynamic and CCrc32Static, each of which acquire the above quatern functions for a numerate of octet. The only conflict between the noise and energizing classes is the CRC table. With the disturbance family the CRC fare and all the functions in the sort are electricity. The business off is uncomplicated. The unchanging league is simpler to use, but the propulsive categorise uses remembering statesman expeditiously because the CRC plateau (1K in filler) is exclusive allocated when needed.

0 comments:

Post a Comment