Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
발표자�소개
• 윤제상�(Jesang�Yoon)�
• (현)�HBSmith�Co�Founder,�CTO�
• (전)�Kanizsa�Lab�CTO�
• (전)�삼성전자�무선사업부�선임연구원�
• 연락처�
• LinkedIn:�jesangyoon�
• GitHub:�@yoonjs2�
• Email:�[email protected]�
• Blog:�https://medium.com/@yoonjs2
S3:�Simple�Storage�Service
• Simple�Storage�Service�
• 클라우드�기반의�Object�File�System(MetaData,�KeyValue,�One�level)�
• 저렴한�비용,�큰�용량,�빠른�접근�
• 고가용성:�최근�장애로�좀…
SPA:�Single�Page�Application
• HTML�기반�Web�Application�
• Ajax�+�Template�=�Dynamic�Contents�
• index.html�단일�페이지에서 동적으로�DOM을�변경하여� 컨텐츠를�보여주는�형태�
• AngularJS,�BackboneJS�등�
• 높은�코드�재�사용성,�데이터�절약,�높은�생산성
S3에서�HTML�호스팅하기
•S3를�Apache�Web�Server�처럼�사용�
• Bucket�Policy:�GetObject�
• Enable�website�hosting�
• S3의�높은�가용성�+�서버관리�부담�Zero
문제:�SPA�Routing
• SPA의�Routing�기능은�URL따라�페이지를�동적으로�생성�
• 동적으로�생성되기�때문에�웹서버�상에�실제�존재하지�않는�페이지임 (Angular에서�#으로�시작하는�주소)�
• html5의�history�기능(html5mode)을�이용할�경우 ‘#’문자�제거�가능�(#은�deprecated�됨)�
• #이�없으면�웹서버에서�존재하지�않는�페이지를�가져오려고�시도한�것으로�간주하여�404�Not�Found를�반환�
• 존재하지�않는�페이지를�존재하는�것처럼�만드는�방법은?
http://foobar.com/#/mypage
http://foobar.com/mypage
/mypage/index.html
200 OK 404 Not Found
/index.html
200 OK
문제:�Eventual�Consistency
• S3는�고가용성을�위해�데이터를�분산�복제�
• 빠른�응답성능을�위해� 데이터가�모든�곳에서�동기화�되기전엔과거�데이터를�읽어올수도�있음(덮어쓰기�PUT,�DELETE)�
• Read�after�Write�Consistency:새로운�Key로�업로드한�데이터는�분산복제가�끝나기�전엔�읽어올수�없음(신규�PUT)�
• 업데이트를�해도�과거�데이터가�다운로드�되는�문제를�해결할�방법은�없을까?(모든�유저가�가급적�최신�컨텐츠를�볼수�있도록)
key:�x key:�x
User�A User�B
덮어쓰기
SPA�Routing�in�S3
• S3의�Redirection�Rule�기능을�사용하여�구현가능�
• 404�Not�Found가�발생할�경우� Error�Page�대신�특정�패턴의�URL로�이동하도록�하는�기능을�응용�
•단점:�#!이�잠시�노출됨�
•주의:�https를�명시적으로�붙여주지�않으면� 앞단에서�https를�강제해도�http로�접속가능하게�됨
Request
S3�RoutingHTTP�301�Redirect
HTML5�mode�(#!�제거)
Eventual�Consistency�회피
• S3의�신규�Key를�가진�Object에�대한�Read�after�Write�Consistency를�이용(처음�업로드�하는�파일들에�대한)�
• S3에서�제공되는�웹�컨텐츠�파일들의�이름에� Timestamp�또는�Hash�값(MD5,�SHA1,�…)을�붙여서� 계속�Key가�바뀌게�함�
• Index.html�같이�Key가�변경되면�안되는�파일들은 CloudFront/CDN�등에서�Purge�또는�Invalidate�시켜 강제로�최신�컨텐츠가�빨리�배포될수�있게�해줌�
•주의:�SPA�처럼�index.html�하나에서�모든�컨텐츠를�불러올수�있는�경우에�유리한�방법� (index.html�하나만�Eventually�Consistency를�신경쓰면�됨)
파일�내용에�대한�MD5�해시값을�계산하여�이용�:�변경사항이�없을�경우�Eventually�Consistent�Read의�장점을�이용가능�
(빠른�컨텐츠�제공)
Bucket간�sync를�응용한�SPA�배포
•AWS�CLI�S3�sync�명령어로�bucket�간�내용�동기화가�가능�
• sync를�이용,�source�bucket의�내용으로�target�bucket의�내용을�채움�
• —delete�옵션으로�source�bucket에�없는�target�bucket의�내용을�청소가능�
• Jenkins/TravisCI�내에서�Git,�Grunt/Gulp와�AWS�CLI�sync를�엮어�완전한�SPA�배포�자동화를�실현�가능
Source bucket Target bucket (Web Hosting Enabled)
퍼블리싱�작업��컨텐츠�최적화�
보안처리�Cache�Breaker�적용
S3 Sync
Purge Cache
주의:�계정간�Bucket�sync의�맹점
• AWS�CLI�S3�명령어는�계정간에�적접�Object를�이동시킬때Object�Metadata의�Canonical�ID를�알아서�변경해주지�않음(이�내용은�문서화가�잘�되어있지�않아�실수할�가능성이�있음)�
• 계정�A의�Object�X를�계정�B의�버켓으로�CLI를�통해�버켓간에�직접�이동시키면�(cp,�mv,�sync)�X의�소유주�Canonical�ID는�여전히�계정�A가�됨�
• 따라서�X는�계정�B에�존재하지만,�계정�B의�Root�ID로도�접근할수�없는�Object가�되어버림(삭제밖에�안됨)�
• 이런�현상을�막으려면�Access�Delegation�을�이용하여�계정�B에서�계정�A의�X를�접근할때�Canonical�ID를�계정�B의�것으로�변경할수�있도록�허용해�줘야�함�
• 자세한�내용:�https://goo.gl/ukQNbs
계정�A 계정�B
S3 Sync
X XCID:�A CID:�A
더�많은�자료는�
페이스북�AWS�KRUG�그룹�
또는�
https://medium.com/@yoonjs2�
에서��
얻어가세요~
Related Documents
