Zcash fi ẹni tí ìnáwó náà jẹ́ fún pamọ̀. Ìdáàbòbò yìí dá ìbéèrè kan tó ń dààmú sílẹ̀: bí nẹ́tàwọ̀n kò bá lè rí àwọn ìnájà ti ó wà lọ́dọ̀ rẹ, báwo ni àpòòwé tìrẹ ṣe máa wá wọn?
Today the answer is blunt. A wallet cannot ask a server which transactions are mine, because that question would reveal exactly what Zcash is trying to hide. So instead the wallet downloads a large amount of data and tests each item locally to see what belongs to it. It works, and it preserves privacy, but it is slow and heavy. This scanning is one of the main reasons wallet syncing can feel sluggish.
Ohun ti o dara julọ ni ọna fun apamọwọ lati beere olupin kan fun gangan data tirẹ, ati gba rẹ, laisi olupin naa kọ ẹkọ ohun ti a beere. Iyẹn jẹ deede kini gbigba alaye ikọkọ pese.
Ìwífún ìsọ̀rọ̀ àdáni jẹ́ ọ̀nà ìdánimọ̀ tí ó ń jé kí oníṣe ka àkọsílẹ̀ kan láti inú ibi-ìpamọ́ àwọn ohun èlò alágbàtà láìfi hàn sí olùrànlọ́wọ́ náà èyí tó kà.
fojú inú wo ilé ìkàwé kan níbi tí o ti lè rí ìwé tóo fẹ́ gbà, ṣùgbọ́n olùtọ̀nà ọ̀hún kò mọ èyí tí wọ́n fún ẹ. ìwọ ni yóò gba ohun náà, àti pé àwọn nǹkan rẹ ò ní jẹ àjèjì sí ẹnikẹ́ni. PIR jẹ́ èrò orí ìṣirò nípa irú ìmọ̀ràn bẹ́ẹ̀, a sì lò ó lórí báńkì èyíkéyìí.
The concept has been studied in cryptography for decades. It was first introduced in 1995 by Chor, Goldreich, Kushilevitz, and Sudan, who described the multiple server approach, and the first single server version followed in 1997 from Kushilevitz and Ostrovsky. It is not something Zcash invented, it is an established field that Zcash is now applying to a real and stubborn problem.
Ọ̀nà méjì ló wà láti kọ́ PIR, ìyàtọ̀ sì ṣe pàtàkì.
Àkọ́kọ́ máa ń lo ọ̀pọ̀lọpọ̀ àwọn servers. Olùgbérè náà rán gbogbo àwọn server lọ láti ṣe ìwádìí, tí yóò sì pa ìdáhùn wọn pọ̀ ní àdúgbò. Kò sí sérver kan ṣoṣo tó rí ohun tí wọ́n béèrè fún yìí mọ́. Èyí jẹ́ òótọ́ o, ṣùgbọ́n ó dá lórí wípé àwọn sérvérì kò bá ara wọn lòdì-kejì, èyí ti ó ṣòro lati ṣèlérí nínú ayé gidi.
Nibi, onibara nlo ohun elo pataki ti a pe ni homomorphic encryption. Eyi jẹ itọsọna to wulo julọ fun awọn imuse gidi nitori ko nilo ọpọlọpọ awọn olupin ti kii ṣe collude .
Homomorphic encryption je iru ifasilẹ ti o jẹ ki olupin kan ṣe iṣiro lori data lakoko ti o wa ni titẹ sii. Olupese naa nmu idahun itọkasi to tọ laisi ri awọn iye ipilẹ.
Eyi ni ero ti o wa lẹhin PIR olupin-ọkan kan ṣe ọna yii. Onibara fẹ nkan nọmba mẹta lati inu atokọ kan. O kọ ibeere kan eyiti, ni ipa, jẹ bẹẹni ti a fi pamọ fun ipo kẹta ati pe ko si ohun elo fun gbogbo awọn ipo miiran. Si olupin naa, ìbéèrè yi nikan ni ariwo alainiye, ko le sọ eyi ti ipo mu bẹẹni.
The server then combines its database with this encrypted query using the special properties of homomorphic encryption, multiplying each stored item by the matching encrypted yes or no and adding the results together. What comes out is a single encrypted package that contains exactly the item the client wanted, and nothing reveals which one it was. The client decrypts that package and reads its item. The server has answered the question without ever knowing the question.
Ẹ̀dà tó lágbára jù, tí a pè ní PIR symmetric, fi ìdánilójú kejì kún un: oníbàárà mọ ohun tó béèrè fún nìkan ni kò sì mọ nǹkankan nípa àwọn àkọsílẹ̀ mìíràn nínú ibi ìpamọ́. Èyí dáàbò bo ibùdó-ìpèsè náà àti onítọ̀hún pàápàá.
Modern single-server schemes are built on lattice cryptography, most commonly the learning with errors assumption. The client's query is a vector of ciphertexts, an encryption of one at the target index and zero elsewhere, and the encryption is additively homomorphic, so the server can add ciphertexts and multiply them by plaintext database entries without decrypting.
Olùgbéejáde náà máa ń ṣe ìwádìí nípa ibi-ìpamọ́ bí matrix, ó sì fi àdìpọ̀ àwọn ohun tí a ti kọ sínú ìwé yìí síbi tó yẹ. Nítorí pé ìbéèrè kò yàtọ̀ sí ariwo àìròtójúṣe, olúgbéeṣé kì í rí ìmọ̀ kankan lórí àmì ọ̀rọ̀ inú rẹ̀.
The historic obstacle has always been cost. Naively, the server must touch every entry in the database for each query, which is expensive in computation, and the ciphertexts are large, which is expensive in bandwidth. Recent research attacks this with preprocessing, schemes such as SimplePIR and FrodoPIR let the server prepare the database ahead of time and hand each client a small hint, pushing much of the work into an offline phase so that live queries become fast. A useful side benefit is that lattice-based constructions are also considered resistant to quantum attacks, which aligns with Zcash's wider move toward post-quantum privacy.
PIR jẹ apakan ti igbiyanju lati ṣe Zcash ni ikọkọ ati iyara lori iwọn.
The wallet scanning bottleneck described earlier is the target. Work at the Valar Group is developing private information retrieval techniques so that a wallet can fetch its own data from a server without the server learning which entries were requested. One concrete application is checking nullifiers privately. A nullifier is a unique marker published when a note is spent, which stops the same funds being spent twice. A wallet often needs to check whether a given nullifier has appeared yet, in other words whether a note is still unspent, and doing that through a server today can leak which note is being asked about. Private information retrieval lets the wallet ask that question without revealing which nullifier it cares about. This sits alongside other scaling work, including Project Tachyon and new node software, aimed at removing the performance limits that hold back private wallets today.
O ṣe pataki lati jẹ olóòótọ́ nípa ipele. Eyi ni iwadi ati imọ-ẹrọ ti nṣiṣe lọwọ, kii ṣe ẹya pari kan, fifunni. Erongba naa wa daradara ati itọsọna ti ṣeto, ṣugbọn ṣiṣe PIR munadoko to fun awọn apamọwọ lojoojumọ lori awọn ẹrọ lasan ni gangan apakan lile ti a ṣiṣẹ lori bayi.