姜春宇-以标准推动开源技术在行业的落地

以标准推动开源技术在行业的落地
展开查看详情

1.զຽ‫ٵ‬വۖ୏რದ๞ࣁᤈӱጱ៧ࣈ ‫ݪل‬ғӾࢵ‫௳מ‬᭗‫מ‬Ꮈᑪᴺ ᘳ֖ғय़හഝದ๞Ԇᓕ ᄍᦖᘏғভชਜ

2.౯ժฎ᧡Ҙ Who are we? य़හഝದ๞ຽ‫ٵ‬വᬰ঩ާտ TC601 ౮ᒈԭଙ11์ ฎӾࢵ᭗‫מ‬ຽ‫ٵ‬۸‫ܐ‬տ CCSA ӥᦡጱӫӱದ๞঩ާտ҅ਤ෰ฎ‫ٽ‬ᘸԾӱ᱾ ‫ݱ‬ӻሾᜓ҅ᦩ‫޾ڦ‬ᥴ٬य़හഝ‫઀ݎ‬ᶎԁጱ᯿य़ᳯ᷌҅զຽ‫ٵ‬വᬰૡ֢ԅᕘଃ҅വۖय़හഝӨਫ֛ᕪၧႮଶᣟ‫ݳ‬ TC601, established in November 2017, is a professional technical committee under CCSA of China Communications Standardization Association. Its purpose is to identify and solve the major problems in big data area, and promote the deep integration of big data and real economy with standard promotion as a link. 1 ਯփࢵਹຽ‫ٵ‬۸ဩ஌̵ဩᥢ޾ොᰒ඲ᒽ҅‫׏‬ᬰԆᓕ᮱ᳪӨտާԏᳵጱԻၞӨဋ᭗Ҕ 2 ୏઀य़හഝᶾऒದ๞޾Ծߝຽ‫ٵ‬Ꮈᑪ҅ᕟᕢտާ୏઀ຽ‫ٵ‬ᒧ‫ݳ‬௔ᦶḵӨᦧ֌ḵᦤૡ֢Ҕ ာ 3 ୏઀හഝ୏න̵හഝၞ᭗̵හഝᩒԾᓕቘᒵᎸᑪ҅୏઀ፘ‫޾ᧃߎى‬ङᦒ๐‫ۓ‬Ҕ ຽ 4 ൫ୌ඲̵Ծ̵਍̵Ꮈ̵አ‫݇ݶو‬ӨጱԻၞଘ‫׏҅ݣ‬ᬰय़හഝӨᤈӱଫአႮଶᣟ‫ݳ‬Ҕ ‫ٵ‬ 5 ൫ୌࢵٖक़य़හഝದ๞޾ԾӱጱԻၞ‫֢҅ݳ‬ᑌຄ݇Өࢵᴬຽ‫ٵ‬۸ૡ̶֢

3.ԅՋԍᵱᥝຽ‫ٵ‬ҘWhy standard is important? ದ๞ெԍᭌҘአಁஉ१य़ So many technologies? It’s really confusing for users. Yahoo!S4 Streaming

4.౯ժࣁ؉ՋԍҘ Link users and developers ຽ‫ٵ‬ Strandard ୏რᐒ‫܄‬ մӱ Open source community Users ‫୏ݻ‬რᬌ‫ف‬մӱጱᵱ࿢ ‫ݻ‬մӱᬌ‫ݢف‬ᶌጱෛದ๞ Collect requirements for developers Help users to choose new available technology

5.ၞᦇᓒጱຽ‫ ٵ‬Stream processing standard ಅզ౯ժࣁଙ҅ᘶ‫ݳ‬ਹմӱ؉ԧᒫӞӻ‫૲ړ‬ୗၞ॒ቘଘ‫ݣ‬ጱຽ‫ٵ‬ So in 2018, we co-work with eight companies to make the first standard for distributed stream processing platform.

6.ၞᦇᓒጱຽ‫ ٵ‬Standard detail ຽ‫ٵ‬ਧԎԧ‫૲ړ‬ୗၞ॒ቘଘ‫ݣ‬ፘ‫ى‬ጱ‫᦯ݷ‬๞᧍҅ᥢਧԧӞӻսᐹጱԾߝଫٍ॓ጱ य़ᔄದ๞ᥝ࿢޾ᶱٍ֛ದ๞೰ຽ̶ The standard defines terminology related to distributed stream processing platform, specifies seven categories of technical requirements and 51 specific technical indicators for an excellent product. च๜‫ۑ‬ᚆ ‫૲ړ‬ୗၞ॒ቘଘ‫ݣ‬ଫٍ॓ጱ॒ቘਫ෸හഝጱच๜ᚆ‫ێ‬ Function Ability to process real time data ّ਻ᚆ‫ێ‬ ‫૲ړ‬ୗၞ॒ቘଘ‫ݣ‬ଫٍ॓ጱӨԆၞᏝկ̵᫫կ݊य़හഝኞாጱّ਻௔ Compatibility Compatible with mainstream hardware, software and big data ecosystem ᓕቘᚆ‫ێ‬ ‫૲ړ‬ୗၞ॒ቘଘ‫ݣ‬ଫٍ॓ጱ੒አಁ‫݋‬অጱᓕቘᚆ‫ێ‬ Management Operational and Maintenance Management Ability ਻Კᚆ‫ێ‬ ‫૲ړ‬ୗၞ॒ቘଘ‫ࣁݣ‬Ꮭկ౲᫫կ୑ଉఘ‫٭‬ӥՖᚆ‫ᦤכ‬හഝӨ๐‫ۓ‬ᑞਧᬩᤈጱᚆ‫ێ‬ Fault tolerance Ensuring stable operation of data and services in case of hardware or software breakdown ಘ઀௔ ‫૲ړ‬ୗၞ॒ቘଘ‫ࣁݣ‬ᕚಘᖽ਻ጱᚆ‫ێ‬ Scalability On-line scaling capability ਞ‫ق‬௔ ‫૲ړ‬ୗၞ॒ቘଘ‫ݣ‬ଫٍ॓ጱਞ‫כق‬ᵑᚆ‫ێ‬ Security Safety and security capability ௔ᚆ ‫૲ړ‬ୗၞ॒ቘଘ‫ݣ‬ଫᘍ੊ጱ௔ᚆ೰ຽᶱ Performance

7. य़හഝຽ‫ٵ‬۸֛ᔮ ౯ժጱٌ՜ຽ‫ٵ‬ਫ᪢ ૡٍᔄຽ‫ٵ‬ හഝ๐‫ۓ‬ᔄ ෛᒈᶱ෸ଧහഝପ̵‫૲ړ‬ୗԪ‫ۓ‬හഝପ̵ၞᦇ ᓒ̵Ꭳᦩࢶᨏ޾ग़ොਞ‫ᦇق‬ᓒᒵӻຽ‫ݚ҅ٵ‬ Ŏ‫૲ړ‬ୗಢ॒ቘଘ‫ݣ‬ Ŏ᭗አ‫ڞٵ‬ Ŏ‫૲ړ‬ୗ‫ړ‬ຉࣳහഝପ Ŏ஄‫מ‬Ծߝ๐‫ۓ‬ क़੤ᦶԧᶎ‫ݻ‬አಁᤈԅහഝ‫ړ‬ຉᥴ٬ොໜጱຽ Ŏහഝᵞ౮ૡٍ ŎᰂᣟᷚഴԾߝ ‫҅ٵ‬ፓ‫ڹ‬૪ᕪ୵౮ದ๞Ծߝᔄຽ‫ٵ‬ᶱ̶ Ŏහഝᓕቘଘ‫ݣ‬ Ŏහഝၞ᭗Ӿஞ Ŏහഝ೵യଘ‫ݣ‬ Ten new standards, such as time series database, distributed Ŏࠟӱฬᚆҁ%,҂‫ړ‬ຉଘ‫ݣ‬ transaction database, stream computing, knowledge map and Ŏ‫૲ړ‬ୗහഝପ multi-party security computing, have been established. At present, Ŏ෸ଧහഝପ 32 technical product standards have been published . Ŏ‫૲ړ‬ୗၞ॒ቘଘ‫ݣ‬ ŏŏ

8.ই֜ḵᦤຽ‫ٵ‬ य़හഝԾߝᚆ‫ᦧێ‬ၥ ᕪᬦଙጱ‫&'҅઀ݎ‬$य़හഝԾߝᚆ‫ێ‬ၥᦧ҅૪ᕪ ౮ԅࢵٖ๦শጱᦧၥ֛ᔮ҅౮ԅ‫ࠟܯ‬ԾߝᎸ‫޾ݎ‬አ ᥴ٬ොໜ አಁᤈԅ Ꭳᦩࢶᨏ ᷚᴾᓕቘ ಁ᯻ᨻᭌࣳጱᷚ‫ݻ‬ຽ̶ After four years of development, DCA big data ‫ړ‬ຉᔄ හഝ೵യ %,ૡٍ ൤ᔱ୚ක product capability evaluation has become an authoritative evaluation system in China. හഝ හഝ හഝ ग़ොਞ‫ق‬ හഝԻഘ හഝᚙභ ᓕቘᔄ ᵞ౮ ᓕቘଘ‫ݣ‬ ᦇᓒ ‫و‬Ձ ‫૲ړ‬ୗ चᏐᚆ‫ێ‬ ௔ᚆ चᏐᔄ ‫૲ړ‬ୗ ‫ړ‬ຉࣳහ ෸ଧහഝ ၞᦇᓒ ‫૲ړ‬ୗԪ ٖਂ ࣋ว੕‫ݻ‬ ಢ॒ቘ ପ ‫ۓ‬හഝପ හഝପ ೰ຽ੕‫ݻ‬ Ծߝ ഝପ

9.ၥᦧᥡ੊ Test observation ᘶ Ⴎ ኪ ᭗ Ӿ Ӿ ‫ܖ‬ ࢥ း Ӿ ᛯӳ ଠ ᛯ ۹ Ꭺ ࢵ ᆂ ྎ ‫ܗ‬ก ࣉ ᐟ ᯡ ‫ܗ‬ ‫ݳ‬ ॠ Ӿ ۹ Gr ᚸෛ ‫מ‬ य़ ኪ սጯ ฃ ᆽ ܏จ ࢵᴨጯᑏӾ ᘶ य़ ො Ӟ ᘶ‫ܖ‬ॠො ૞ ॠ Ղහ ۸ ᗑ ᅉ ࿥ ਜኼ܏‫܏ڹ‬ᛪ ‫ظ‬ ฃ ᙑ හၵਪ᫫ Ղ࿞๔ႆ ૤ ᅾ ow ᐟ ᦔ܏ ԯ හ ᑀ ‫ړظ‬ Ấ ‫ݢ‬ ԅሾ ‫݌‬᯾ଶᝁ‫ي‬ మ ᭗ ւ ೰ ႎশ‫מ‬ᰂ ࿤ ௤ ᕪ໼ ፄ Պ ᭗ ᵞ ᭗හਜၹ౬᭗ හ ฬ ᬡ ฬᄤ‫ࢵמ‬ හၒ࣡௏ ๩ ኂ ing ᒽ ԯӣ ‫ل‬ ഝ  ஑ᅩ ഠ ኞ Ꮈ አ ӱ ੢ ᭗ ௳‫מ‬ ฬ จ ᕋ ᑀ ‫ڥ‬ ‫מ‬ ౮ ᬡഝ ‫מ‬ አ ഝ ᫫ ‫ܗ‬ ᜮ ᴬ ᧍ IO ‫ݪ‬ ‫ل‬ ಅ ௳ ‫ݪ‬ ಢ॒ቘଘ‫ݣ‬௔ᚆ ‫ړ‬ຉࣳහഝପ௔ᚆ ಢ॒ቘଘ‫ݣ‬चᏐ ‫ړ‬ຉࣳහഝପचᏐ හഝᵞ౮ හഝᓕቘ ࠟ‫ۓ‬ฬᚆ හഝ೵യ ෸ଧහഝପ ‫૲ړ‬ୗԪ‫ۓ‬හഝପ አಁᤈԅහഝ‫ړ‬ຉ 134ӻ ፓ‫ڹ‬૪ᕪํԾߝ᭗ᬦᦧၥ҅ਠ౮ၥᦶහᰁӻ̶ At present, 83 products have passed the evaluation and 134 tests have been completed.

10.ၥᦧᥡ੊ Test observation 100TB 10000ᜓᅩ 300ᜓ ๋य़चᏐᚆ‫ێ‬ᵞᗭᥢཛྷ ๋य़ጱ௔ᚆၥᦶᥢཛྷ ‫ق‬ቖᶾ‫ض‬හഝᥢཛྷ ௔ᚆ๋य़ၥᦶහഝ7% ᴨ᯾ਠ౮ᜓᅩಢ ܏ԅ᭗ᬦԧᜓᅩᇔቘ๢௔ᚆᦧၥ ॒ቘଘ‫ݣ‬चᏐᚆ‫ᦧێ‬ၥ ᴨ᯾᭗ᬦӻԯ๐‫࢏ۓ‬௔ᚆᦧၥ ෛ܏ӣ᭗ᬦԧᜓᅩᇔቘ๢௔ᚆᦧၥ ๋ࡅཻጱᦧၥᶱፓ ‫ق‬ᚆᭌಋഭᤈག ಢ॒ቘଘ‫ݣ‬चᏐ ಢ॒ቘଘ‫ݣ‬௔ᚆ ࢵ‫݌‬ ‫ړ‬ຉࣳහഝପचᏐ จሾ හഝᵞ౮ ෸ଧහഝପ ӾᑏۖᝁᎸ හഝ೵യ ᚸᦔԯ ‫૲ړ‬ୗԪ‫ۓ‬හഝପ හഝᓕቘ Ӿ‫ي‬ ‫ړ‬ຉࣳහഝପ௔ᚆ ኪ‫מ‬ԯ‫ݪل‬ አಁᤈԅ‫ړ‬ຉ ࠟ‫ۓ‬ฬᚆ ᴨ᯾

11. ‫ݸ‬ᖅ‫ ֢ݳ‬Future Collaboration Ø Ө)OLQNᐒ‫૲ݎ֢ݳ܄‬ၞ॒ቘଘ‫ݣ‬ਫ᪢ጮጼԡ Ø Publish Steam Processing Platform White Paper with Flink Community Ø ୏઀ᶎ‫ݻ‬մӱአಁጱದ๞ङᦒ Ø Organize Technical Tutorial for Enterprise Users Ø ᬰᤈ‫૲ړ‬ୗၞ॒ቘଘ‫ݣ‬௔ᚆຽ‫ٵ‬ጱ‫ګ‬ਧ Ø Making Benchmark for Performance Ø ‫ۖސ‬ၞ॒ቘଘ‫ݣ‬ᒫӞ᫪ၥᦧ Ø Start the test for Stream Processing Platform

12.