Haeba bohlale ba maiketsetso bo nkuoa e le leeto ho tloha A ho ea ho B, cloud computing service ke boema-fofane kapa seteishene sa terene se lebelo le holimo, 'me edge computing ke tekesi kapa baesekele e arolelanoang.Computing ea Edge e haufi le lehlakore la batho, lintho, kapa mehloli ea data.E amohela sethala se bulehileng se kopanyang polokelo, khomphutha, phihlello ea marang-rang, le bokhoni ba mantlha ba ts'ebeliso ho fana ka lits'ebeletso ho basebelisi ba tikolohong eo.Ha ho bapisoa le lits'ebeletso tsa komporo ea leru tse kentsoeng bohareng, komporo ea moeli e rarolla mathata a kang latency e telele le sephethephethe se phahameng sa convergence, ho fana ka ts'ehetso e ntle bakeng sa lits'ebeletso tsa nako ea nnete le tse hlokang bandwidth.
Mollo oa ChatGPT o thehile leqhubu le lecha la nts'etsopele ea AI, ho potlakisa ho teba ha AI libakeng tse ngata tsa kopo tse kang indasteri, mabenkele, matlo a bohlale, metse e bohlale, joalo-joalo Palo e kholo ea data e hloka ho bolokoa le ho baloa ho pheletso ea kopo, 'me ho itšetleha ka leru feela ha ho sa khona ho fihlela tlhoko ea nnete, komporo ea moeli e ntlafatsa kilometara ea ho qetela ea lits'ebetso tsa AI.Tlas'a leano la naha la ho nts'etsapele moruo oa dijithale ka matla, khomphutha ea maru ea China e kene nakong ea nts'etsopele e kenyellelitsoeng, tlhoko ea komporo ea moeli e eketsehile, 'me ho kopanngoa ha maru le pheletso e fetohile tataiso ea bohlokoa ea ho iphetola ha lintho nakong e tlang.
'Maraka oa komporo oa Edge o holisa 36.1% CAGR nakong ea lilemo tse hlano tse tlang
Indasteri ea k'homphieutha e kene boemong ba tsoelo-pele e tsitsitseng, joalo ka ha ho pakoa ke ho fetoha ha bafani ba litšebeletso butle-butle, boholo ba 'maraka bo ntseng bo eketseha, le katoloso e eketsehileng ea libaka tsa kopo.Mabapi le boholo ba mmaraka, tlhaiso-leseling e tsoang tlalehong ea ho latela ea IDC e bonts'a hore boholo ba mmaraka oa li-server tsa khale tsa China bo fihlile ho US $ 3.31 bilione ka 2021, mme boholo ba mmaraka oa li-server tsa China tse lebelletsoeng ho hola ka kholo ea selemo le selemo. sekhahla sa 22.2% ho tloha 2020 ho isa 2025. Sullivan o bolela esale pele hore boholo ba mmaraka ba komporo ea moeli Chaena bo lebelletsoe ho fihla ho RMB 250.9 bilione ka 2027, le CAGR ea 36.1% ho tloha 2023 ho isa 2027.
Indasteri ea eco-ea khomphutha ea Edge e ea atleha
Computing ea Edge hajoale e mothating oa pele oa ho phatloha, 'me meeli ea khoebo khoebong ea indasteri e batla e le lerootho.Bakeng sa barekisi ka bomong, hoa hlokahala ho nahana ka ho kopanngoa le maemo a khoebo, hape hoa hlokahala ho ba le bokhoni ba ho ikamahanya le liphetoho tsa maemo a khoebo ho tloha boemong ba theknoloji, hape hoa hlokahala ho netefatsa hore ho na le tekanyo e phahameng ea ho lumellana le lisebelisoa tsa hardware, hammoho le bokhoni ba boenjiniere ba ho etsa merero ea mobu.
Ketane ea indasteri ea komporo e arotsoe ka barekisi ba chip, barekisi ba algorithm, baetsi ba lisebelisoa tsa Hardware, le bafani ba tharollo.Barekisi ba li-chip hangata ba nts'etsapele li-chips tsa arithmetic ho tloha pheletsong ho ea pheletsong ho isa lehlakoreng la maru, 'me ntle le li-chips tse lehlakoreng, ba boetse ba theha likarete tsa ho potlakisa le sethala sa nts'etsopele ea software.Barekisi ba algorithm ba nka li-algorithms tsa pono ea khomphutha e le mantlha ho aha li-algorithms tse akaretsang kapa tse hlophisitsoeng, hape ho na le likhoebo tse hahang litsi tsa algorithm kapa lithupelo le li-platform tsa push.Barekisi ba lisebelisoa ba tsetela ka mafolofolo lihlahisoa tsa komporo ea mantlha, 'me mofuta oa lihlahisoa tsa komporo ea moeli o lula o ntlafatsoa, butle-butle ho etsa lihlahisoa tse ngata tsa komporo ho tloha ho chip ho isa mochining oohle.Bafani ba tharollo ba fana ka tharollo ea software kapa software-hardware-integrated bakeng sa liindasteri tse itseng.
Lisebelisoa tsa indasteri ea komporo ea Edge lia potlakisa
Sebakeng sa motse o bohlale
Hajoale tlhahlobo e felletseng ea thepa ea litoropo e sebelisoa hangata ka mokhoa oa tlhahlobo ea matsoho, 'me mokhoa oa ho hlahloba ka letsoho o na le mathata a litšenyehelo tse jang nako e ngata le tse boima ba basebetsi, ho itšetleha ka batho ka bomong, ho fana ka tšireletso e fokolang le khafetsa ea tlhahlobo, le boleng bo tlase. taolo.Ka nako e ts'oanang ts'ebetso ea tlhahlobo e tlaleha palo e kholo ea data, empa lisebelisoa tsena tsa data ha lia fetoloa hore e be thepa ea data bakeng sa ho matlafatsa khoebo.Ka ho sebelisa theknoloji ea AI maemong a tlhahlobo ea mehala, khoebo e thehile koloi ea tlhahlobo e bohlale ea taolo ea litoropo ea AI, e sebelisang mahlale a kang Marang-rang a Lintho, komporo ea maru, li-algorithms tsa AI, 'me e nkile lisebelisoa tsa botsebi joalo ka lik'hamera tsa litlhaloso tse phahameng, ho- lipontšo tsa boto, le li-server tsa AI tse lehlakoreng, 'me li kopanya mokhoa oa ho hlahloba "mochine o bohlale + mochine o bohlale + thuso ea basebetsi".E khothaletsa phetoho ea puso ea litoropo ho tloha ho bohlale ba basebetsi ho ea ho bohlale ba mochini, ho tloha ho kahlolo e matla ho ea ho tlhahlobo ea data, le ho tloha karabelong e sa sebetseng ho isa ho sibollo e sebetsang.
Sebakeng sa setša sa kaho se bohlale
Litharollo tsa sebaka sa kaho tse bohlale tse thehiloeng ho komporo li sebelisa kopanyo e tebileng ea theknoloji ea AI mosebetsing oa khale oa ho lekola polokeho ea indasteri ea kaho, ka ho beha setsi sa tlhahlobo sa AI sebakeng sa kaho, ho phethela lipatlisiso tse ikemetseng le nts'etsopele ea li-algorithms tsa pono tsa AI tse thehiloeng ho video e bohlale. thekenoloji ya analytics, ho lemoha ka nako e tletseng diketsahalo tse lokelang ho lemohuwa (mohlala, ho lemoha hore na motho o roalwe helmete kapa che), ho fana ka basebetsi, tikoloho, tshireletso le boitsebahatso ba maemo a kotsi a polokeho le ditshebeletso tsa dikgopotso tsa alamo, le ho nka bohato ba ho Itsebahatsa ba sa bolokehang. lintlha, tlhokomelo e bohlale ea AI, ho boloka litšenyehelo tsa basebetsi, ho fihlela litlhoko tsa taolo ea polokeho ea thepa ea basebetsi le thepa ea libaka tsa kaho.
Lefapheng la lipalangoang tse bohlale
Mehaho ea li-cloud-side-end e fetohile paradigm ea motheo bakeng sa ho tsamaisoa ha likopo indastering e bohlale ea lipalangoang, 'me lehlakore la maru le ikarabellang bakeng sa tsamaiso e bohareng le karolo ea ts'ebetso ea data, lehlakore le ka thōko haholo-holo le fana ka tlhahlobo ea lintlha tse ka thōko le qeto ea khomphutha. -ho etsa ts'ebetso, le lehlakore la ho qetela le ikarabellang haholo bakeng sa pokello ea data ea khoebo.
Maemong a ikhethileng joalo ka khokahano ea 'mila oa koloi, mateano a holographic, ho khanna ka boiketsetso, le sephethephethe sa terene, ho na le lisebelisoa tse ngata tse fapaneng tse fihlelletsoeng, mme lisebelisoa tsena li hloka taolo ea phihlello, taolo ea ho tsoa, ts'ebetso ea alamo, le ts'ebetso le ts'ebetso ea tlhokomelo.Computing ea Edge e ka arola le ho hlola, ea fetola e kholo ho ba e nyane, ea fana ka mesebetsi ea phetolo ea protocol ea cross-layer, ea fihlela phihlello e kopaneng le e tsitsitseng, esita le taolo e kopanetsoeng ea data e fapaneng.
Lefapheng la tlhahiso ea indasteri
Boemo ba Ntlafatso ea Ts'ebetso ea Tlhahiso: Hajoale, palo e kholo ea lits'ebetso tsa tlhahiso e sa sebetseng e lekantsoe ke ho se phethehe ha data, 'me katleho ea lisebelisoa ka kakaretso le lipalo tse ling tsa lintlha tsa index li batla li le bohlasoa, li etsa hore ho be thata ho li sebelisa bakeng sa ts'ebetso e nepahetseng.Edge komporong sethaleng e thehiloeng thepa tlhahisoleseding ea mohlala ho finyella semantic boemo ba tlhahiso ea tsamaiso ea rapameng puisano le paatsepama puisano, e thehiloeng sebele nako ya data phallo sebetsa mochine ho aggregate le ho sekaseka palo e khōlō ea data tšimo nako ea sebele, ho finyella mohlala-based tlhahiso line. fusion ea tlhaiso-leseling ea mehloli e mengata, ho fana ka ts'ehetso e matla ea data bakeng sa ho etsa liqeto ka har'a sistimi ea tlhahiso ea discrete.
Equipment Predictive Maintenance Scenario: Paballo ea thepa ea indasteri e arotsoe ka mefuta e meraro: tokiso ea ho lokisa, tokiso e thibelang, le tlhokomelo e boletsoeng esale pele.Tlhokomelo ea tsosoloso ke ea ts'ebetso ea ex-post facto, tlhokomelo ea thibelo, le tlhokomelo ea pele-pele ke ea tlhokomelo ea mehleng ea khale, ea pele e itšetlehile ka nako, ts'ebetso ea lisebelisoa, maemo a sebaka, le lintlha tse ling bakeng sa tlhokomelo ea kamehla ea thepa, ho feta kapa ka tlaase ho motho. phihlelo, ea ho qetela ka ho bokella lintlha tsa sensor, ho hlahloba nako ea sebele ea boemo ba ts'ebetso ea lisebelisoa, ho itšetlehile ka mohlala oa indasteri oa tlhahlobo ea data, le ho bolela esale pele ka nepo hore na ho hlōleha ho etsahala neng.
Maemo a tlhahlobo ea boleng ba indasteri: sebaka sa tlhahlobo ea pono ea indasteri ke mofuta oa pele oa tloaelo oa tlhahlobo ea othomathike (AOI) sebakeng sa tlhahlobo ea boleng, empa nts'etsopele ea AOI ho fihlela hajoale, ka ho lemoha liphoso tse ngata le maemo a mang a rarahaneng, ka lebaka la mefokolo ea mefuta e fapaneng. ea mefuta e, tšobotsi e ntšitsoeng ha e ea fella, ditharabololo feto-fetoha futsanehile extensibility, mola tlhahiso e ntjhafatswa kgafetsa, ho falla ha algorithm ha e tenyetsehe, 'me lintlha tse ling, tsamaiso ea setso AOI e bile thata ho kopana le ntshetsopele ya litlhoko tsa mola tlhahiso.Ka hona, sethala sa algorithm sa tlhahlobo ea boleng ba indasteri ea AI se emeloang ke thuto e tebileng + ho ithuta sampole e nyane butle-butle e nkela sebaka sa tloaelo sa tlhahlobo ea pono, mme sethala sa tlhahlobo ea boleng ba indasteri ea AI se fetile methating e 'meli ea li-algorithms tsa khale tsa ho ithuta le li-algorithms tsa tlhahlobo e tebileng ea ho ithuta.
Nako ea poso: Oct-08-2023