On-premise bụ usoro ebe ụlọ ọrụ na-ewulite ma na-arụkwa akụrụngwa IT ha ozugbo n’ime ụlọ kọmputa ma ọ bụ sava nke ha, kama iji igwe ojii dị n’èzí. Ịnakwere usoro On-premise na-enye ụlọ ọrụ uru nke ịchịkwa nchekwa data na sistemụ ha ozugbo.
Akwụkwọ a na-akọwa n’ụzọ teknụzụ uru ndị bụ isi a pụrụ ịtụ anya ha mgbe a nabatara Clevi On-Premise, yana gburugburu kọmpụta e kewapụrụ nke ọma ma chekwaa maka ụdị Clevi-X MDLM(Massive Multimodal Language Model).
1. Uru ndị a na-enweta mgbe a nabatara On-Premise
Ịkwalite nnwere onwe data na nchekwa ya ruo n’ókè kachasị
- A na-arụ usoro niile—site n’itinye data, nyocha, ọrụ, ruo nkwado ndabere—n’onwe ya n’ime netwọkụ mechiri emechi nke ụlọ ọrụ. → A na-ebelata ihe ize ndụ nke ịpụ data n’èzí, a na-akwalitekwa ikike nke ụlọ ọrụ nwere n’ebe data ya dị na nchekwa ozi nkeonwe.
- A na-ekewapụ data nwere mmetụta, dị ka Fine Dataset na dataset nkeonwe, n’ime mpaghara nchekwa dị iche → Nke a na-eme ka ịchịkwa nnweta data na ime nyocha ya dị mfe.
Nhazi akụrụngwa nwere arụmọrụ dị elu na nnweta dị elu
- A na-ahazi sava GPU/CPU n’ụdị klasta n’akụkụ nyocha na ịgbakọ → A na-ahazi ya nke ọma maka nnukwu mgbakọ AI, nhazi data ozugbo, na nyocha yiri ibe ya.
- A na-etinye nhazi HA(High Availability) → Ọ na-enye ohere ịrụ ọrụ nke ọma na-enweghị nkwụsị ọrụ ọbụna mgbe nsogbu mere.
Njikwa data na-agbanwe agbanwe na ike ịgbasawanye
- A na-ekewapụ klasta nchekwa data jikọtara ọnụ na ebe nchekwa nkwado ndabere oyi → Nke a na-eme ka a na-achịkwa data ozugbo na data a na-echekwa ogologo oge nke ọma.
- A haziri mpaghara nyocha, ọrụ, na njikwa data n’ụdị modul, ya mere ọ dị mfe ịgbasawanye na ịrụzi ya → Nke a na-enye ohere ịzaghachi ngwa ngwa mgbanwe ọrụ, mmụba data, na iwebata ọrụ ọhụrụ.
SVCE(Clevi Safe Virtural Compute Environment)
- Nkwalite nchekwa → Na-ekewapụ ọrụ ọ bụla na nnọkọ onye ọrụ nke ọma, na-egbochi ihe ize ndụ nke ịrụ koodu ọjọọ, ịpụpụ data na imebi sistemụ.
- Mgbanwe na ikike ịgbasawanye → A na-ahazi ọnụ ọgụgụ instances na oke akụrụngwa n'ụzọ dị ike dabere na mkpa, ma na-enye gburugburu ahaziri maka ụdị ọrụ AI dị iche iche (ederede, onyonyo, koodu wdg.).
- Ịrụ ọrụ nke ọma → Site na njikwa akụrụngwa na nhazi oge akpaghị aka, ọ na-eme ka ojiji akụrụngwa akụrụngwa dị elu, ma na-akwado imepụta instance ọhụrụ na iweghachi n'ụdị gara aga ngwa ngwa mgbe mmejọ mere.
2. Gbasara gburugburu kọmputa dịpụrụ adịpụ echekwara maka ụdị Clevi-X MDLM
Clevi-X MDLM (M Double LM - Massive Multimodal Language Model) bụ ụdị AI multimodal nke ọgbọ ọhụrụ e mere iji mụta gburugburu na ụzọ, nhazi na usoro mmadụ si eji kọmputa, ma kwado ọrụ kọmputa n'ezie.
Gghe gburugburu akpa dị fechaa dabere na CLK (Clevi Linux Kernel) ka e wee hụ na mwakpo nje na ịnweta n'ụzọ ọjọọ yana ọrụ ọ bụla gafere ikike bụ ihe na-agaghị ekwe omume.
- Ọbụna n'ime gburugburu igwe ojii ebe ọtụtụ ndị ọrụ a na-amaghị aha na-abanye maka ebumnuche dị iche iche, SVCE(Safe Virtual Compute Environment) na-ekewapụ nnọkọ na ọrụ ọ bụla nke ọma, na-ahụ na nchekwa dị n'agbanyeghị mwakpo ma ọ bụ mbọ ịbanye nke ndị ọrụ ọjọọ.
- Ebe ọrụ niile dị ka ịnweta ịntanetị, ịgbakọ, ịrụ koodu, inyocha nsonaazụ na itinye ma ọ bụ wepụ faịlụ na-eme n'ime oghere dịpụrụ adịpụ echekwara, a na-ekwe nkwa ntụkwasị obi nke data mmepụta ụdị ahụ na ikike ịrụ ọrụ.
- Ọ na-enye gburugburu echekwara pụọ n'ihe iyi egwu dị iche iche, gụnyere mbanye anataghị ikike, ịpụpụ data, mwakpo na-eji ụdị ahụ na mwakpo a na-achọ ụdị ahụ.
Ajụjụ na arịrịọ ngosi: [email protected]
Copyright© 2025 Clevi Inc. All rights reserved.