Ukuba i-artificial intelligence ithathwa njengohambo olusuka ku-A ukuya ku-B, inkonzo ye-cloud computing sisikhululo seenqwelo-moya okanye isikhululo sikaloliwe esinesantya esiphezulu, kwaye i-edge computing yiteksi okanye ibhayisekile ekwabelwana ngayo.I-Edge computing isondele kwicala labantu, izinto, okanye imithombo yedatha.Yamkela iqonga elivulekileyo elidibanisa ukugcinwa, ukubala, ukufikelela kwinethiwekhi, kunye nezakhono eziphambili zesicelo ukubonelela ngeenkonzo kubasebenzisi kwindawo ekufutshane.Xa kuthelekiswa neenkonzo ze-computing zefu ezibekwe kwindawo ephakathi, i-edge computing isombulula iingxaki ezifana ne-latency ende kunye ne-high convergence traffic, ukubonelela ngenkxaso engcono kwixesha langempela kunye neenkonzo ezifuna i-bandwidth.
Umlilo we-ChatGPT uqalise i-wave entsha yophuhliso lwe-AI, ukukhawuleza ukucwiliswa kwe-AI kwiindawo ezininzi zesicelo ezifana nezoshishino, ukuthengisa, amakhaya ahlakaniphile, izixeko ezihlakaniphile, njl njl. Inani elikhulu ledatha kufuneka ligcinwe kwaye libalwe kwi-computing. ukuphela kwesicelo, kwaye ukuthembela kwilifu lodwa akusenako ukuhlangabezana nemfuno yokwenene, i-edge computing iphucula ikhilomitha yokugqibela yezicelo ze-AI.Phantsi komgaqo-nkqubo wesizwe wokuphuhlisa ngamandla uqoqosho lwedijithali, i-computing yelifu yaseTshayina ingene kwixesha lophuhliso olubandakanyayo, imfuno yekhompyuter esekupheleni iye yanda, kwaye ukudityaniswa komphetho welifu kunye nokuphela kube lukhokelo olubalulekileyo kwikamva.
Imakethi yekhompyuter ye-Edge iza kukhula nge-36.1% CAGR kule minyaka mihlanu izayo
Ishishini le-edge computing lingene kwinqanaba lophuhliso oluzinzileyo, njengoko kungqinwa kukwahlukahlukana ngokuthe ngcembe kwababoneleli beenkonzo balo, ukwanda kobungakanani bemarike, kunye nokwandiswa ngakumbi kweendawo zezicelo.Ngokumalunga nobungakanani bemakethi, idatha evela kwingxelo yokulandela umkhondo ye-IDC ibonisa ukuba ubungakanani bemarike bubonke beeseva zekhompyuter e-China bufikelele kwi-3.31 yeebhiliyoni zeedola ngo-2021, kwaye ubungakanani bemarike bubonke beeseva zekhompyuter e-China kulindeleke ukuba bukhule ngokukhula ngonyaka. izinga le-22.2% ukusuka ku-2020 ukuya ku-2025. Uqikelelo lwe-Sullivan ubungakanani bemarike ye-edge computing e-China kulindeleke ukuba ifikelele kwi-RMB 250.9 yebhiliyoni ngo-2027, kunye ne-CAGR ye-36.1% ukusuka kwi-2023 ukuya kwi-2027.
IEdge computing eco-industry iyaphumelela
I-Edge computing okwangoku ikwinqanaba lokuqala loqhambuko, kwaye imida yeshishini kumxokelelwano woshishino ayintsonkothanga.Kubathengisi ngabanye, kuyimfuneko ukuqwalasela ukudityaniswa kunye neemeko zoshishino, kwaye kuyimfuneko ukuba ube nokukwazi ukulungelelanisa utshintsho kwiimeko zoshishino ukusuka kwinqanaba lobugcisa, kwaye kuyafuneka ukuba kuqinisekiswe ukuba kukho iqondo eliphezulu. ukuhambelana nezixhobo zehardware, kunye nobuchule bobunjineli bomhlaba weeprojekthi.
Umda wekhonkco loshishino lwekhompuyutha lohlulwe lwaba ngabathengisi betshiphu, abathengisi be-algorithm, abavelisi bezixhobo zehardware, kunye nababoneleli besisombululo.Abathengisi beChip ikakhulu baphuhlisa iitshiphusi ze-arithmetic ukusuka kwicala ukuya kumda ukuya kwicala lelifu, kwaye ukongeza kwi-edge-side chips, baphuhlisa amakhadi okukhawulezisa kunye nenkxaso yamaqonga ophuhliso lwesoftware.Abathengisi be-algorithm bathatha ii-algorithms zombono wekhompyuter njengondoqo wokwakha ii-algorithms eziqhelekileyo okanye ezilungiselelweyo, kwaye kukho amashishini akha ii-algorithms ze-algorithms okanye uqeqesho kunye namaqonga otyhala.Abathengisi bezixhobo batyala imali kwiimveliso zekhompyuter, kwaye imo yeemveliso zekhompyuter zihlala zityetyiswa, ngokuthe ngcembe zenza isitaki esipheleleyo seemveliso zekhompyuter ukusuka kwitshiphu ukuya kumatshini wonke.Ababoneleli besisombululo babonelela ngesoftware okanye izisombululo ezidityanisiweyo zesoftware kumashishini athile.
Usetyenziso lweshishini lwekhompuyutha luyakhawuleza
Kwintsimi yesixeko esihlakaniphile
Uhlolo olubanzi lwepropathi yasezidolophini ngoku ngokuqhelekileyo lusetyenziswa ngendlela yokuhlola ngesandla, kwaye imowudi yokuhlola ngesandla ineengxaki zokuchitha ixesha elininzi kunye neendleko eziphezulu zabasebenzi, ukuxhomekeka kwenkqubo kumntu ngamnye, ukhuseleko olulambathayo kunye nokuhlolwa rhoqo, kunye nomgangatho ophantsi. ulawulo.Ngexesha elifanayo inkqubo yokuhlola irekhode inani elikhulu ledatha, kodwa ezi zixhobo zedatha azizange zitshintshwe zibe yimpahla yedatha yokuxhobisa ishishini.Ngokusebenzisa itekhnoloji ye-AI kwiimeko zokuhlola eziphathwayo, ishishini lenze isithuthi sokuhlola esikrelekrele se-AI solawulo lwasezidolophini, esamkela itekhnoloji efana ne-Intanethi yeZinto, i-computing yamafu, i-algorithms ye-AI, kwaye iphethe izixhobo zobuchwephesha ezinjengeekhamera ezinenkcazo ephezulu, kwi- iziboniso zebhodi, kunye nabancedisi becala le-AI, kwaye idibanisa indlela yokuhlola "inkqubo ekrelekrele + umatshini okrelekrele + uncedo lwabasebenzi".Ikhuthaza inguqu yolawulo lwasezidolophini ukusuka kubukrelekrele babasebenzi ukuya kubukrelekrele boomatshini, ukusuka kubungqina obunamandla ukuya kuhlalutyo lwedatha, kunye nokusuka kwimpendulo engenzi nto ukuya ekubhaqweni okusebenzayo.
Kwintsimi yesiza sokwakha esikrelekrele
Izisombululo zendawo yokwakha ekrelekrele esekwe kwikhompyuter zisebenzisa ukudityaniswa okunzulu kwetekhnoloji ye-AI kumsebenzi wokubeka iliso wokhuseleko lweshishini lolwakhiwo lwemveli, ngokubeka isiphelo sohlalutyo lwe-AI kwindawo yokwakha, ukugqiba uphando oluzimeleyo kunye nophuhliso lwe-algorithms ye-AI ebonakalayo esekwe kwividiyo ekrelekrele. itekhnoloji ye-analytics, ukubonwa kwexesha elizeleyo leziganeko eziza kubhaqwa (umzekelo, ukukhangela ukuba unxibe okanye unganxibi isigcina-ntloko), ukubonelela ngabasebenzi, okusingqongileyo, ukhuseleko kunye nezinye iinkonzo zokuchonga umngcipheko wokhuseleko kunye neenkonzo zokukhunjuzwa kwe-alam, kunye nokuthatha inyathelo lokuqala lokuchonga abangakhuselekanga. izinto, AI ukugada obukrelekrele, ukonga iindleko zabasebenzi, ukuhlangabezana abasebenzi kunye neemfuno zolawulo lokhuseleko lwepropathi kwiindawo zokwakha.
Kwintsimi yothutho olukrelekrele
I-design-side-end architecture ibe yi-paradigm esisiseko yokuthunyelwa kwezicelo kwishishini lezothutho ezihlakaniphile, kunye necala lelifu elijongene nolawulo oluphakathi kunye nenxalenye yokucubungula idatha, icala lomda libonelela ngokuhlalutya kwedatha yecala kunye nesigqibo sokubala. -ukwenza ukusetyenzwa, kunye necala lokugqibela elinoxanduva ikakhulu lokuqokelelwa kwedatha yeshishini.
Kwiimeko ezithile ezifana nokulungelelaniswa kwemoto-indlela, ukunqumla kwe-holographic, ukuqhuba ngokuzenzekelayo, kunye nokuhamba ngomzila kaloliwe, kukho inani elikhulu lezixhobo ezingafaniyo ezifunyenweyo, kwaye ezi zixhobo zifuna ulawulo lokufikelela, ulawulo lokuphuma, ukusetyenzwa kwe-alamu, kunye nokusebenza kunye nokulungiswa kokugcinwa.I-Edge computing inokwahlula kwaye yoyise, ijike ibe nkulu ibe yincinci, ibonelele ngemisebenzi yokuguqulwa kweprotocol enqamlezayo, ukufikelela ngokudibeneyo kunye nokufikelela okuzinzileyo, kunye nolawulo olusebenzisanayo lwedatha eyahlukileyo.
Kwintsimi yokwenziwa kwemizi-mveliso
Imeko yokuPhuculwa kweNkqubo yeMveliso: Okwangoku, inani elikhulu leenkqubo zokuvelisa ezidityanisiweyo zisikelwe umda kukungaphelelanga kwedatha, kunye nokusebenza kakuhle kwezixhobo kunye nezinye izibalo zedatha yesalathiso zihamba kancinci, zisenza kube nzima ukusebenzisela ukusebenza kakuhle.Edge computing platform esekelwe kwisixhobo ulwazi imodeli ukuphumeza inkqubo yokuvelisa inqanaba semantic unxibelelwano oluthe tye kunye nonxibelelwano nkqo, esekelwe kwi-real-time data flow processing indlela yokudibanisa kunye nokuhlalutya inani elikhulu ledatha yexesha lokwenyani kwintsimi, ukufezekisa umgca wemveliso osekelwe kwimodeli. Ukuhlanganiswa kolwazi lwemithombo yedatha emininzi, ukubonelela ngenkxaso yedatha enamandla yokwenza izigqibo kwinkqubo yokuvelisa.
I-Equipment Predictive Maintenance Scenario: Ukugcinwa kwezixhobo zemizi-mveliso kwahlulwe kwaba ziindidi ezintathu: ulungiso olulungisayo, ugcino lothintelo, kunye nokugcinwa kwangaphambili.Ulondolozo lokubuyisela kwimeko yesiqhelo lolokugcinwa kwe-ex post facto, ulondolozo lothintelo, kunye nolondolozo oluqikelelwayo lololondolozo lwangaphambili, olokuqala lusekwe kwixesha, ukusebenza kwesixhobo, iimeko zesiza, kunye nezinye izinto zokugcinwa rhoqo kwezixhobo, ngaphezulu okanye ngaphantsi kusekwe ebantwini. amava, okokugqibela ngokuqokelelwa kwedatha ye-sensor, ukubeka iliso ngexesha langempela lemeko yokusebenza yezixhobo, ngokusekelwe kwimodeli yoshishino yohlalutyo lwedatha, kunye nokuqikelela ngokuchanekileyo xa kwenzeka ukungaphumeleli.
Imeko yokuhlolwa komgangatho weshishini: intsimi yokuhlola umbono woshishino luhlobo lokuqala lokuhlola oluzenzekelayo oluzenzekelayo (AOI) kwintsimi yokuhlola umgangatho, kodwa uphuhliso lwe-AOI ukuza kuthi ga ngoku, kwiindawo ezininzi zokubona iziphene kunye nezinye iimeko ezinzima, ngenxa yeziphene ezahlukeneyo. kweentlobo, ukutsalwa kwesici akuphelelanga, i-algorithms eguquguqukayo ukwandiswa kakubi, umgca wemveliso uhlaziywa rhoqo, ukufuduka kwe-algorithm ayiguqukiyo, kunye nezinye izinto, inkqubo ye-AOI yendabuko ibe nzima ukuhlangabezana nophuhliso lweemfuno zomgca wemveliso.Ke ngoko, iqonga le-algorithm yokuhlola umgangatho weshishini elimelwe kukufunda okunzulu + isampula encinci yokufunda ngokuthe ngcembe ithatha indawo yeskim sokuhlola okubonwayo, kwaye iqonga lokuhlola umgangatho weshishini le-AI liye ladlula kwizigaba ezibini ze-algorithms yokufunda koomatshini kunye ne-algorithms yokuhlola nzulu yokufunda.
Ixesha lokuposa: Oct-08-2023