I kēia lā, e kamaʻilio mākou e pili ana i ka maikaʻi o ka mālama ʻana i ka ʻikepili i loko o kahi honua kahi e hoʻopuka ai nā pūnaewele ʻelima, nā genome scanners a me nā kaʻa kaʻa kaʻa ponoʻī i ka ʻikepili i kēlā me kēia lā ma mua o nā kānaka āpau i hana ʻia ma mua o ka hoʻololi ʻoihana.
Ke hana nei ko mākou honua i nā ʻike hou aʻe. ʻO kekahi hapa o ia mea he pōkole a nalowale koke e like me ka hōʻiliʻili ʻana. Pono e mālama ʻia kahi ʻē aʻe, a ua hoʻolālā ʻia kekahi "no nā kenekulia" - ʻo ia ka mea a mākou e ʻike ai mai kēia manawa. Ke holo nei ka ʻike i nā kikowaena ʻikepili i ka wikiwiki a hiki i kekahi ala hou, kekahi ʻenehana i hoʻolālā ʻia e hoʻokō i kēia "koi" pau ʻole e lilo koke.
ʻO 40 mau makahiki o ka hoʻomohala ʻana i nā ʻōnaehana mālama puʻupuʻu
ʻO ka waihona pūnaewele mua ma ke ʻano a mākou i kamaʻāina ai i ʻike ʻia i ka makahiki 1980. Ua ʻike ka nui o ʻoukou i ka NFS (Network File System), AFS (Andrew File System) a i ʻole Coda. He ʻumi makahiki ma hope mai, ua loli ke ʻano a me ka ʻenehana, a ua hāʻawi ʻia nā ʻōnaehana waihona puʻupuʻu i nā ʻōnaehana mālama puʻupuʻu ma muli o GPFS (General Parallel File System), CFS (Clustered File Systems) a me StorNext. Ua hoʻohana ʻia ka mālama ʻana o ka papa hana maʻamau ma ke ʻano he kumu, ma luna o ia mea i hana ʻia ai kahi ʻōnaehana faila me ka hoʻohana ʻana i kahi papa polokalamu. Hoʻohana mau ʻia kēia mau mea a me nā ʻano like, noho i kā lākou niche a koi nui ʻia.
I ka hoʻomaka ʻana o ka millennium, ua loli iki ka paradigm i hoʻokaʻawale ʻia, a ua lawe nā ʻōnaehana me SN (Shared-Nothing) i nā kūlana alakaʻi. Ua hoʻololi ʻia mai kahi waihona puʻupuʻu i ka mālama ʻana ma nā nodes pākahi, ʻo ia hoʻi, he mau kikowaena maʻamau me ka lako polokalamu e hāʻawi ana i kahi mālama pono; Ma ia mau loina, e ʻōlelo ʻia, kūkulu ʻia ʻo HDFS (Hadoop Distributed File System) a me GFS (Global File System).
Ma kahi kokoke i ka 2010s, ua hoʻomaka ka ʻike ʻia ʻana o nā manaʻo e pili ana i nā ʻōnaehana mālama i hoʻolaha ʻia i loko o nā huahana kalepa piha, e like me VMware vSAN, Dell EMC Isilon a me kā mākou . Ma hope o nā paepae i ʻōlelo ʻia, ʻaʻohe kaiāulu o nā mea hoihoi, akā nā mea kūʻai kikoʻī i kuleana no ka hana, kākoʻo, a me ka lawelawe ʻana o ka huahana a hōʻoia i kona hoʻomohala hou ʻana. ʻO ia mau hoʻonā i makemake nui ʻia ma kekahi mau wahi.
Mea lawelawe kelepona
Malia paha kekahi o nā mea kūʻai mua loa o nā ʻōnaehana mālama i hoʻolaha ʻia ʻo ia nā mea lawelawe kelepona. Hōʻike ke kiʻikuhi i nā pūʻulu o nā noi e hana i ka nui o ka ʻikepili. ʻO OSS (Operations Support Systems), MSS (Management Support Services) a me BSS (Business Support Systems) e hōʻike ana i ʻekolu papa lako polokalamu pono e hoʻolako i ka lawelawe i ka poʻe kākau inoa, hōʻike kālā i ka mea hāʻawi a me ke kākoʻo hana i nā ʻenekini hana.
ʻO ka manawa pinepine, ua hui pū ʻia ka ʻikepili o kēia mau papa me kekahi, a i mea e pale aku ai i ka hōʻiliʻili ʻana o nā kope pono ʻole, hoʻohana ʻia ka mālama ʻana, e hōʻiliʻili i ka nui o ka ʻike e hele mai ana mai ka pūnaewele hana. Hoʻohui ʻia nā waihona i loko o kahi pūnāwai maʻamau, i loaʻa i nā lawelawe āpau.
Hōʻike kā mākou helu ʻana i ka hoʻololi ʻana mai nā ʻōnaehana mālama maʻamau i ka pale ʻana i nā ʻōnaehana mālama e hiki ai iā ʻoe ke mālama i ka 70% o ka waihona kālā wale nō ma ka haʻalele ʻana i nā ʻōnaehana mālama hi-end i hoʻolaʻa ʻia a me ka hoʻohana ʻana i nā kikowaena hoʻolālā maʻamau maʻamau (maʻamau x86), e hana pū me nā loea kūikawā. lako polokalamu. Ua hoʻomaka mua nā mea hoʻohana kelepona e kūʻai i ia mau hopena i nā nui nui. Ma keʻano kūikawā, ua hoʻohana nā mea hana Lūkini i nā huahana mai Huawei no nā makahiki ʻeono.
ʻAe, ʻaʻole hiki ke hoʻopau ʻia kekahi mau hana me ka hoʻohana ʻana i nā ʻōnaehana puʻunaue. No ka laʻana, me ka hoʻonui ʻana i nā koi hana a i ʻole ka hoʻohālikelike ʻana me nā protocol kahiko. Akā ma ka liʻiliʻi o 70% o ka ʻikepili i hoʻoponopono ʻia e ka mea hoʻokele hiki ke loaʻa i loko o kahi loko i hoʻokaʻawale ʻia.
ʻĀpana panakō
I loko o kēlā me kēia panakō he nui nā ʻōnaehana IT ʻokoʻa, e hoʻomaka ana mai ka hana ʻana a hoʻopau me kahi ʻōnaehana panakō automated. Hoʻohana pū kēia ʻōnaehana me ka nui o ka ʻike, ʻoiai ʻo ka hapa nui o nā hana ʻaʻole pono ka hoʻonui ʻana i ka hana a me ka hilinaʻi o nā ʻōnaehana mālama, no ka laʻana, hoʻomohala, hoʻāʻo, automation o nā kaʻina keʻena, a me nā mea ʻē aʻe. aka, i kela makahiki keia makahiki, ua emi iho ka waiwai. Eia hou, i kēia hihia,ʻaʻohe mea maʻalahi i ka hoʻohanaʻana i nā kumuwaiwai pūnaewele mālama, aʻo ka hana o ia mea i heluʻia ma muli o ka ukana kiʻekiʻe.
I ka hoʻohana ʻana i nā ʻōnaehana mālama ʻia, hiki ke hoʻololi ʻia ko lākou mau nodes, ʻo ia hoʻi nā kikowaena maʻamau, i kēlā me kēia manawa, no ka laʻana, i loko o kahi mahiʻai server a hoʻohana ʻia ma ke ʻano he papa helu.
Nā loko ʻikepili
Hōʻike ka kiʻi ma luna i kahi papa inoa o nā mea kūʻai lawelawe maʻamau . Hiki i kēia mau lawelawe e-government (no ka laʻana, "Government Services"), nā ʻoihana digitalized, nā ʻoihana kālā, a me nā mea ʻē aʻe. Pono lākou a pau e hana me ka nui o nā ʻike heterogeneous.
ʻAʻole pono ka hoʻohana ʻana i nā ʻōnaehana mālama maʻamau no ka hoʻoponopono ʻana i ia mau pilikia, no ka mea, pono ia i ke komo ʻana i ka hana kiʻekiʻe no ka pale ʻana i nā waihona a me ke komo maʻamau i nā hale waihona puke o nā palapala scan i mālama ʻia ma ke ʻano he mau mea. Eia kekahi laʻana, hiki ke hoʻopili ʻia kahi ʻōnaehana kauoha ma o ka puka pūnaewele. No ka hoʻokō ʻana i kēia mau mea āpau ma kahi kahua mālama maʻamau, pono ʻoe i kahi hoʻonohonoho nui o nā mea hana no nā hana like ʻole. Hiki ke uhi ʻia kahi ʻōnaehana mālama honua ākea i nā hana a pau i helu mua ʻia: pono ʻoe e hana i kekahi mau loko me nā ʻano waihona ʻokoʻa i loko.
Nā mea hana i nā ʻike hou
Ke ulu nei ka nui o ka ʻike i mālama ʻia ma ka honua ma kahi o 30% i kēlā me kēia makahiki. He nūhou maikaʻi kēia no nā mea kūʻai aku mālama, akā he aha a ʻo ia ke kumu nui o kēia ʻikepili?
He ʻumi makahiki i hala aku nei, ua lilo nā ʻoihana pūnaewele i nā mea hoʻomohala, pono kēia i ka hana ʻana i kahi helu nui o nā algorithms hou, nā hāmeʻa hāmeʻa, a me nā mea ʻē aʻe. ʻO ka mea mua, ʻo ia ka ʻike kapua. I kēia manawa, ma kahi o 70% o nā hui e hoʻohana i nā lawelawe kapuaʻi ma kekahi ala a i ʻole. Hiki i kēia mau ʻōnaehana leka uila, kope kope a me nā mea virtualized ʻē aʻe.
ʻO ka lua o ka mea hoʻokele ka lima o ka hanauna. ʻO kēia nā wikiwiki hou a me nā helu hoʻoili data hou. Wahi a kā mākou wānana, ʻo ka hoʻolaha ākea o 5G e alakaʻi i kahi hāʻule o ka noi no nā kāleka hoʻomanaʻo flash. No ka nui o ka hoʻomanaʻo i loko o ke kelepona, e pau ana ia, a inā he 100-megabit channel ka gadget, ʻaʻohe pono e mālama i nā kiʻi ma ka ʻāina.
ʻO ke kolu o nā kumu kumu e ulu ai ka noi no nā ʻōnaehana mālama e pili ana i ka hoʻomohala wikiwiki ʻana o ka naʻauao artificial, ka hoʻololi ʻana i ka ʻikepili ʻikepili nui a me ke ʻano i ka automation universal o nā mea āpau.
ʻO kahi hiʻohiʻona o ka "new traffic" ʻo ia . Pono mākou e mālama i kēia ʻikepili me ka wehewehe ʻole ʻana i kona ʻano ma kekahi ʻano. Pono ia no ka heluhelu hope. No ka laʻana, no ka hoʻoholo ʻana i ka nui o ka hōʻaiʻē i loaʻa, e nānā kahi ʻōnaehana helu panakō i nā kiʻi āu e kau ai ma nā pūnaewele kaiapuni, e hoʻoholo inā hele pinepine ʻoe i ke kai a i nā hale ʻaina, a ma ka manawa like e aʻo i nā unuhi mai kāu mau palapala lapaʻau i loaʻa. ia ia. ʻO kēia mauʻikepili, ma kekahiʻaoʻao, he laulā, akā ma kekahiʻaoʻao, nele i ka homogeneity.
Ka moana o ka ʻikepili i kūkulu ʻole ʻia
He aha nā pilikia e pili ana i ka puka ʻana mai o nā "ʻikepili hou"? ʻO ka mea mua i waena o lākou, ʻoiaʻiʻo, ʻo ka nui o ka ʻike a me ka manawa i manaʻo ʻia o kona mālama ʻana. ʻO kahi kaʻa kaʻa kaʻa kaʻa ʻole o kēia wā wale nō e hoʻopuka i ka 60 terabytes o ka ʻikepili i kēlā me kēia lā mai kāna mau mea ʻike a me nā mīkini. No ka hoʻomohala ʻana i nā algorithm neʻe hou, pono e hana ʻia kēia ʻike i loko o ka lā hoʻokahi, i ʻole e hoʻomaka e hōʻiliʻili. I ka manawa like, pono e mālama ʻia no ka manawa lōʻihi - mau makahiki. Ma laila wale nō e hiki ai i ka wā e hiki mai ana ke huki i nā hopena ma muli o nā laʻana analytical nui.
Hoʻokahi mea no ka deciphering genetic sequences e hoʻopuka ma kahi o 6 TB i kēlā me kēia lā. A ʻo ka ʻikepili i hōʻiliʻili ʻia me kāna kōkua ʻaʻole ia e hōʻike i ka holoi ʻana, ʻo ia hoʻi, hypothetically, pono e mālama mau ʻia.
ʻO ka hope, ʻo nā ʻupena o ka hanauna ʻelima. Ma kahi o ka ʻike maoli i hoʻouna ʻia, ʻo ia pūnaewele ponoʻī he mea hoʻomohala nui o ka ʻikepili: nā moʻolelo hana, nā moʻolelo kelepona, nā hopena waena o nā pilina mīkini-me-machine, etc.
Pono kēia mau mea i ka hoʻomohala ʻana i nā ala hou a me nā algorithms no ka mālama ʻana a me ka hoʻoili ʻana i ka ʻike. A ke puka mai nei ia mau ʻano.
Nā ʻenehana o ke au hou
ʻEkolu mau pūʻulu hoʻonā i hoʻolālā ʻia no ka hoʻokō ʻana i nā koi hou no nā ʻōnaehana mālama ʻike: ka hoʻokomo ʻana i ka naʻauao artificial, ka hoʻomohala ʻenehana o ka media waihona a me nā mea hou i ke kahua o ka hoʻolālā ʻōnaehana. E hoʻomaka kākou me AI.
I nā hoʻonā Huawei hou, hoʻohana ʻia ka naʻauao artificial ma ke kiʻekiʻe o ka waihona ponoʻī, i lako ʻia me kahi kaʻina AI e hiki ai i ka ʻōnaehana ke nānā kūʻokoʻa i kona kūlana a wānana i nā hemahema. Inā pili ka ʻōnaehana mālama i kahi kapuaʻi lawelawe i loaʻa nā mana helu helu nui, hiki i ka naʻauao artificial ke hana i nā ʻike hou aʻe a hoʻonui i ka pololei o kāna mau kuhiakau.
Ma kahi o nā hemahema, hiki i kēlā AI ke wānana i ka ukana kiʻekiʻe e hiki mai ana a me ka manawa i koe a hiki i ka pau ʻana o ka hiki. ʻAe kēia iā ʻoe e hoʻonui i ka hana a hoʻonui i ka ʻōnaehana ma mua o ka hiki ʻana o nā hanana makemake ʻole.
I kēia manawa e pili ana i ka ulu ʻana o ka waihona waihona. Ua hana ʻia nā pahu uila mua me ka ʻenehana SLC (Single-Level Cell). ʻO nā mea i hoʻokumu ʻia ma luna o ia mea he wikiwiki, hilinaʻi, paʻa, akā he liʻiliʻi liʻiliʻi a he pipiʻi loa. Ua hoʻokō ʻia ka ulu ʻana o ka leo a me ka hōʻemi ʻana i ke kumukūʻai ma o kekahi mau ʻae ʻenehana, no laila ua hoʻemi ʻia ka wikiwiki, hilinaʻi a me ke ola lawelawe o nā drive. Eia nō naʻe, ʻaʻole i hoʻopilikia ke ʻano i nā ʻōnaehana mālama ponoʻī iā lākou iho, ma muli o nā ʻano hoʻopunipuni like ʻole, ua lilo i mea ʻoi aku ka maikaʻi a ʻoi aku ka hilinaʻi.
Akā no ke aha ʻoe i makemake ai i nā ʻōnaehana mālama All-Flash? ʻAʻole i lawa ka hoʻololi wale ʻana i nā HDD kahiko i kahi ʻōnaehana hana me nā SSD hou o ke ʻano kumu like? Ua koi ʻia kēia i mea e hoʻohana pono ai i nā kumuwaiwai āpau o nā ʻōnaehana solid-state hou, kahi hiki ʻole i nā ʻōnaehana kahiko.
ʻO Huawei, no ka laʻana, ua hoʻomohala i nā ʻenehana he nui e hoʻoponopono i kēia pilikia, ʻo kekahi o ia , ka mea i hiki ai ke hoʻonui i nā pilina "disk-controller" e like me ka hiki.
ʻO ka ʻike naʻauao i hiki ai ke hoʻokaʻawale i ka ʻikepili i loko o nā kahawai he nui a hoʻopaʻa ʻia i kekahi mau hanana makemake ʻole, e like me (kākau hoʻonui). I ka manawa like, nā algorithms hoʻihoʻi hou, ʻoi aku , hoʻonui i ka wikiwiki o ke kūkulu houʻana, e ho'ēmi ana i kona manawa i nā mea liʻiliʻi loa.
ʻO ka hāʻule ʻole, ka nui, ka ʻohi ʻōpala - ʻaʻole pili hou kēia mau mea i ka hana o ka ʻōnaehana mālama ma muli o nā hoʻololi kūikawā i nā mea hoʻokele.
A ke hoʻomākaukau nei hoʻi nā waihona ʻikepili poloka e hui . E hoʻomanaʻo mākou i ka hana maʻamau no ka hoʻonohonoho ʻana i ka ʻike ʻikepili i hana e like me kēia: ua komo ka mea hana i ka mea hoʻokele RAID ma o ka pahi ʻo PCI Express. ʻO ia hoʻi, ua launa pū me nā diski mechanical ma o SCSI a i ʻole SAS. ʻO ka hoʻohana ʻana i ka NVMe ma ke kua i hoʻolalelale nui i ke kaʻina holoʻokoʻa, akā hoʻokahi drawback: pono e pili pono nā drive i ka mea hana i mea e hāʻawi pololei ai i ka hoʻomanaʻo.
ʻO ka pae aʻe o ka hoʻomohala ʻenehana a mākou e ʻike nei i kēia manawa ʻo ia ka hoʻohana ʻana o NVMe-oF (NVMe over Fabrics). No nā ʻenehana poloka Huawei, ua kākoʻo mua lākou iā FC-NVMe (NVMe ma luna o Fiber Channel), a ʻo NVMe ma RoCE (RDMA ma luna o Converged Ethernet) ma ke ala. Hoʻohana maikaʻi ʻia nā hiʻohiʻona hoʻāʻo; aia kekahi mau mahina i koe ma mua o kā lākou hōʻike kūhelu. E hoʻomaopopo e ʻike ʻia kēia mau mea āpau i nā ʻōnaehana puʻupuʻu, kahi e makemake nui ai ka "Ethernet lossless".
ʻO kahi ala ʻē aʻe e hoʻomaikaʻi ai i ka hana o ka mālama ʻana i hoʻohele ʻia ʻo ka haʻalele piha ʻana i ka hoʻohālikelike ʻikepili. ʻAʻole hoʻohana hou nā hopena Huawei i n kope, e like me ka RAID 1 maʻamau, a hoʻololi loa i ka (Erasure coding). Hoʻohālikelike kahi pūʻolo makemakika kūikawā i nā poloka mana i kekahi manawa, e ʻae iā ʻoe e hoʻihoʻi i ka ʻikepili waena i ka wā o ka nalowale.
ʻO ka hoʻokaʻawale ʻana a me ka hoʻohana ʻana i ke kaomi ʻana he mea pono. Inā i loko o nā ʻōnaehana mālama maʻamau e kaupalena ʻia mākou e ka helu o nā kaʻina hana i hoʻokomo ʻia i loko o nā mea hoʻokele, a laila ma nā ʻōnaehana mālama scalable horizontally, loaʻa i kēlā me kēia node nā mea āpau e pono ai: disks, memory, processors and interconnect. Ua lawa kēia mau kumuwaiwai e hōʻoia i ka liʻiliʻi o ka hopena o ka unuhi ʻana a me ke kaomi ʻana i ka hana.
A e pili ana i nā ʻano hana hoʻomaʻamaʻa. Ma aneʻi ua hiki ke hoʻemi i ka ukana ma nā kaʻina hana waena me ke kōkua o nā chips i hoʻolaʻa ʻia (a i ʻole nā poloka i hoʻolaʻa ʻia i loko o ka kaʻina hana ponoʻī), e pāʻani ana i kahi kuleana. (TCP/IP Offload Engine) a i ʻole ka lawe ʻana i nā hana makemakika o EC, deduplication a me ka hoʻoemi.
Hoʻokomo ʻia nā ala hou no ka mālama ʻana i ka ʻikepili i loko o kahi hoʻolālā i hoʻokaʻawale ʻia. Loaʻa i nā ʻōnaehana kikowaena kikowaena kahi hale hana kikowaena i hoʻopili ʻia ma o Fiber Channel i me ka nui o na papa. ʻO nā hemahema o kēia ala ʻo ia ka paʻakikī o ka scaling a me ka hōʻoia ʻana i kahi pae o ka lawelawe (ma ke ʻano o ka hana a i ʻole latency). Hoʻohana nā ʻōnaehana Hyperconverged i nā pūʻali like no ka mālama ʻana a me ka hoʻoili ʻana i ka ʻike. Hāʻawi kēia i ka palena palena ʻole no ka hoʻonui ʻana, akā e pili ana i nā kumukūʻai kiʻekiʻe no ka mālama pono ʻana i ka ʻikepili.
ʻAʻole e like me nā mea ʻelua ma luna nei, ʻo kahi hoʻolālā hoʻokaʻawale ʻia ka hoʻokaʻawale ʻana i ka ʻōnaehana i kahi lole hoʻopili a me kahi ʻōnaehana hoʻopaʻa ʻana. Hāʻawi kēia i nā pōmaikaʻi o nā hale kiʻi ʻelua a hiki i ka scaling palena ʻole o ka mea wale nō i nele i ka hana.
Mai ka hoʻohui ʻana i ka hui ʻana
ʻO kahi hana maʻamau, ʻo ka pilina o ia mea i ulu wale i nā makahiki he 15 i hala iho nei, ʻo ia ka pono e hāʻawi i ka manawa like i ka mālama ʻana i ka poloka, ke komo ʻana i nā faila, ke komo ʻana i nā mea, ka hana ʻana o kahi māla ʻikepili nui, a pēlā aku. no ka laʻana, he ʻōnaehana hoʻihoʻi ma ka tape magnetic.
I ka pae mua, ʻo ka hoʻokele wale ʻana o kēia mau lawelawe i hiki ke hoʻohui. Ua hoʻopili ʻia nā ʻōnaehana mālama ʻikepili heterogeneous i kekahi polokalamu kūikawā, kahi i puʻunaue ai ka luna hoʻomalu i nā kumuwaiwai mai nā loko i loaʻa. Akā, no ka mea he ʻokoʻa nā lako o kēia mau loko, ʻaʻole hiki ke neʻe i waena o lākou. Ma kahi kiʻekiʻe o ka hoʻohui ʻana, ua kū ka hōʻuluʻulu ma ka pae ʻīpuka. Inā loaʻa ka kaʻana like ʻana i nā faila, hiki ke lawelawe ʻia ma o nā protocols like ʻole.
ʻO ke ʻano hoʻohālikelike ʻoi loa i loaʻa iā mākou i kēia manawa e pili ana i ka hana ʻana i kahi ʻōnaehana hybrid universal. ʻO kā mākou pono e lilo ai . Hoʻohana ka ʻike honua i nā kumuwaiwai like ʻole, hoʻokaʻawale ʻia i loko o nā loko like ʻole, akā ʻae ʻia no ka neʻe ʻana. Hiki ke hana i kēia a pau ma o ka console hoʻokele hoʻokahi. Ma kēia ala, ua hiki iā mākou ke hoʻokō i ka manaʻo o "hoʻokahi kikowaena data - hoʻokahi ʻōnaehana mālama."
ʻO ke kumukūʻai o ka mālama ʻana i ka ʻike i kēia manawa e hoʻoholo ai i nā hoʻoholo hale kūkulu. A ʻoiai hiki ke mālama pono ʻia i mua, i kēia lā ke kūkākūkā nei mākou i ka mālama "ola" me ke komo ʻana, no laila pono e noʻonoʻo ʻia ka hana. ʻO kekahi waiwai koʻikoʻi o nā ʻōnaehana puʻupuʻu e hiki mai ana ka hui ʻana. Ma hope o nā mea a pau, ʻaʻohe mea makemake e loaʻa i nā ʻōnaehana like ʻole e hoʻomalu ʻia mai nā consoles like ʻole. Hoʻopili ʻia kēia mau hiʻohiʻona āpau i ka moʻo hou o nā huahana Huawei .
Pūnaehana mālama nui o ka hanauna hou
Hoʻokō ʻo OceanStor Pacific i ʻeono mau ʻeiwa mau koi hilinaʻi (99,9999%) a hiki ke hoʻohana ʻia e hana i nā kikowaena ʻikepili papa HyperMetro. Me ka mamao ma waena o ʻelua mau kikowaena data a hiki i 100 km, hōʻike nā ʻōnaehana i kahi latency hou o 2 ms, kahi e hiki ai ke kūkulu ma luna o kā lākou kumu i nā hopena kū'ē i ka pōʻino, me ka poʻe me nā kikowaena quorum.
Hōʻike nā huahana moʻo hou i ka versatility protocol. I kēia manawa, kākoʻo ʻo OceanStor 100D i ke komo ʻana i ka poloka, ke komo ʻana i nā mea a me ke komo ʻana o Hadoop. E hoʻokō ʻia ke komo waihona i ka wā e hiki mai ana. ʻAʻohe pono e mālama i nā kope he nui o ka ʻikepili inā hiki ke hoʻopuka ʻia ma o nā protocols like ʻole.
Me he mea lā, he aha ka pili o ka manaʻo o kahi "ʻenehana nalowale" me nā ʻōnaehana mālama? ʻO ka ʻoiaʻiʻo, ua kūkulu ʻia nā ʻōnaehana mālama ʻikepili ma ke kumu o kahi pūnaewele wikiwiki e kākoʻo ana i nā algorithms kūpono a me ka mīkini RoCE. ʻO ka ʻōnaehana naʻauao artificial i kākoʻo ʻia e kā mākou hoʻololi e kōkua i ka hoʻonui ʻana i ka wikiwiki o ka pūnaewele a hoʻemi i ka latency. . ʻO ka loaʻa ʻana o ka hana mālama i ka wā e hoʻāla ai i ka AI Fabric hiki ke hiki i 20%.
He aha ka mea hou o OceanStor Pacific node waihona? Aia ka 5U form factor solution he 120 drives a hiki ke hoʻololi i ʻekolu mau nodes maʻamau, e hāʻawi ana i ʻoi aku ma mua o ʻelua mau waihona kālā ma kahi o ka rack space. Ma ka mālama ʻole ʻana i nā kope, piʻi nui ka pono o nā drive (a hiki i +92%).
Ua maʻa mākou i ka ʻoiaʻiʻo ʻo ka waihona i wehewehe ʻia i ka polokalamu he polokalamu kūikawā i hoʻokomo ʻia ma kahi kikowaena maʻamau. Akā i kēia manawa, no ka hoʻokō ʻana i nā palena maikaʻi loa, pono nō hoʻi kēia hoʻonā hale kūkulu i nā nodes kūikawā. Loaʻa iā ia nā kikowaena ʻelua e pili ana i nā kaʻina hana ARM e hoʻokele ana i kahi ʻano o nā pahu ʻekolu iniha.
ʻAʻole kūpono kēia mau kikowaena no nā hoʻonā hyperconverged. ʻO ka mea mua, he liʻiliʻi nā noi no ARM, a ʻo ka lua, paʻakikī ka mālama ʻana i ke kaulike ukana. Manaʻo mākou e neʻe i kahi waihona hoʻokaʻawale: kahi pūʻulu computing, i hōʻike ʻia e nā kikowaena maʻamau a i ʻole nā rack server, e hana ʻokoʻa, akā pili ʻia i nā node waihona ʻo OceanStor Pacific, nāna e hana i kā lākou mau hana pololei. A ua hoapono ia ia iho.
No ka laʻana, e lawe mākou i kahi hoʻonā mālama ʻikepili nui maʻamau me kahi ʻōnaehana hyperconverged e noho ana i nā pahu kikowaena 15. Inā ʻoe e puʻunaue i ka ukana ma waena o nā kikowaena computing kaʻawale a me nā node mālama ʻana o OceanStor Pacific, e hoʻokaʻawale ana iā lākou mai kekahi i kekahi, e ʻoki ʻia ka helu o nā pahu i makemake ʻia! Hoʻemi kēia i nā kumukūʻai hana o ke kikowaena data a hōʻemi i ka nui o ke kumukūʻai o ka ʻona. I kahi honua kahi e ulu nei ka nui o ka ʻike i mālama ʻia e 30% i kēlā me kēia makahiki, ʻaʻole i hoʻolei ʻia kēlā mau pōmaikaʻi.
***
Hiki iā ʻoe ke loaʻa i ka ʻike hou aku e pili ana i nā hoʻonā Huawei a me kā lākou mau hiʻohiʻona noi ma kā mākou a i ʻole ma ke kamaʻilio pololei ʻana i nā luna ʻoihana.
Source: www.habr.com