Design and Development of a Provenance Capture Platform for Data Science
Ben Sterrett - Sitecore Wars IV: A new way to deliver media - SUGCON
1. Organized by the Community, for the Community.
SITECORE WARS IV: A NEW
HOPE WAYTO DELIVER
MEDIA
Ben Sterrett
2. Organized by the Community, for the Community.SUGCON NORTH AMERICA 2015 2
BEN STERRETT? WHO?
3. Organized by the Community, for the Community.
ABOUT ME
• I went to Clemson
University (GOTIGERS)
• I work for Hedgehog
Development (GO
HEDGEHOGS)
• I love to hack on new
technology
• I have three crazy
daughters
• I like long walks on the
beach
• I like getting caught in
rain
• I hate piña coladas
SUGCON NORTH AMERICA 2015 3
4. Organized by the Community, for the Community.
ABOUT ME
• I have a background in
high performance /
high throughput
computing
• I frequently work with
node and RoR in my
spare time
• I once broke a window
with a baseball and I
don’t know if it was on
purpose
• One time I told a PM
that I would allow him
to Photoshop text into
his images
SUGCON NORTH AMERICA 2015 4
5. Organized by the Community, for the Community.SUGCON NORTH AMERICA 2015 5
UGH, FINE. WHY AREYOU
HERE?
6. Organized by the Community, for the Community.
I LOVETHE INTERNET
SUGCON NORTH AMERICA 2015 6
7. Organized by the Community, for the Community.
I LOVETHE INTERNET
SUGCON NORTH AMERICA 2015 7
Pic related: It’s a bunny
with a sad.
… and slow internets give me a sad.
8. Organized by the Community, for the Community.SUGCON NORTH AMERICA 2015 8
YOU HAVEN’T ANSWERED
THE QUESTION
9. Organized by the Community, for the Community.SUGCON NORTH AMERICA 2015 9
I WANTTOTALK ABOUTTHE
MEDIA LIBRARY
10. Organized by the Community, for the Community.
SITECORE DRIVES CUSTOMER EXPERIENCE
SUGCON NORTH AMERICA 2015 10
… it says it right up front.
11. Organized by the Community, for the Community.
SITECORE DRIVES CUSTOMER EXPERIENCE
• GroundbreakingVersatility
• StatisticalVision IntoYour Site
• Second-To-None Authorship Experiences
• Actionable Metrics to help improve business
SUGCON NORTH AMERICA 2015 11
12. Organized by the Community, for the Community.
SITECORE DRIVES CUSTOMER EXPERIENCE
• Sitecore is Un-opinionated
• It is a Blank Canvas
• It is a web development platform
• It is an advice-giving platform
• Does not care about how you do UX
SUGCON NORTH AMERICA 2015 12
… without getting in your way
13. Organized by the Community, for the Community.SUGCON NORTH AMERICA 2015 13
OF COURSE SITECORE
DOESN’T DO UX, HOW
COULD IT?
14. Organized by the Community, for the Community.
ANSWER: SPEED
• We are a part of a global community
• What works in the US might not work elsewhere
• The rest of Sitecore is ready out-of-the-box to handle a
global marketplace
• The media library needs to reflect this
SUGCON NORTH AMERICA 2015 14
…Lightning-fast, neck-breaking speed
15. Organized by the Community, for the Community.SUGCON NORTH AMERICA 2015 15
WHY ARE WE STILL HOSTING
OUR OWN MEDIA?
16. Organized by the Community, for the Community.
HOW MEDIA EXISTSTODAY
SUGCON NORTH AMERICA 2015 16
17. Organized by the Community, for the Community.
HOW MEDIA EXISTSTODAY
Pros
• Easily curatable content
• Securable
• Reliable
• Keeps DBA salaries high
Cons
• ASHX handlers can be slow
• Databases get gigantic and hugely
difficult to maintain
• Transformations take up valuable
server time
• Web server is busy handling media,
not other web requests
• File synchronization can be an
issue for web farm environments
• The web server exists in one place
geographically
• Not “cloud friendly”
SUGCON NORTH AMERICA 2015 17
18. Organized by the Community, for the Community.SUGCON NORTH AMERICA 2015 18
YES, BEN,THIS ISWHY WE
CREATED CONTENT
DELIVERY NETWORKS
19. Organized by the Community, for the Community.
CONTENT DELIVERY NETWORKS
SUGCON NORTH AMERICA 2015 19
20. Organized by the Community, for the Community.
SITECORE ON CDNS
• No out-of-the-box solution
• Sitecore CDN Connector Module
• Set MediaPrefixUrl to the CDN Url
• AWS Elastic Load Balancer + CloudFront
SUGCON NORTH AMERICA 2015 20
… should be the standard for today
21. Organized by the Community, for the Community.
DOESTHIS SOLVE ALL OF OUR CONS?
Pros
• Easily curatable content
• Securable
• Reliable
• Keeps DBA salaries high
Cons
• ASHX handlers can be slow
• Databases get gigantic and hugely
difficult to maintain
• Transformations take up valuable
server time
• Web server is busy handling media,
not other web requests
• File synchronization can be an
issue for web farm environments
• The web server exists in one place
geographically
• Not “cloud friendly”
SUGCON NORTH AMERICA 2015 21
22. Organized by the Community, for the Community.SUGCON NORTH AMERICA 2015 22
WHY DON’T WE PUT OUR
ASSETS SOMEWHERETHAT
SPECIALIZES IN ASSETS?
23. Organized by the Community, for the Community.
PEOPLE ARE DOINGTHIS ALREADY
SUGCON NORTH AMERICA 2015 23
24. Organized by the Community, for the Community.
WHERE MY WORK COMES IN
SUGCON NORTH AMERICA 2015 24
… initial focus is on images
=+
25. Organized by the Community, for the Community.SUGCON NORTH AMERICA 2015 25
WOAHTHERE, COWBOY.
IMGIX? S3?
26. Organized by the Community, for the Community.
INTRODUCING IMGIX
• Image Manipulation
over HTTP
• CDN
• Giant Server Farms
• Massive Array of
Transformations
SUGCON NORTH AMERICA 2015 26
27. Organized by the Community, for the Community.
INTRODUCING S3
• Cloud-Based Storage
From Amazon
• Structure Provided by
Keys
• Extensive APIs and SDKs
• Can cheaply store large
amounts of data
• Not a CDN
SUGCON NORTH AMERICA 2015 27
28. Organized by the Community, for the Community.SUGCON NORTH AMERICA 2015 28
OH, OKAY. CONTINUE.
29. Organized by the Community, for the Community.
PARTS OFTHE PROCESS
1. Upload media blob data to S3 instead of Sitecore
(keeping image item)
2. Extend Field Renderer
3. Redirect Media Item requests
SUGCON NORTH AMERICA 2015 29
30. Organized by the Community, for the Community.SUGCON NORTH AMERICA 2015 30
STOP,YOU NEEDTO SHOW
METHIS NOW.
31. Organized by the Community, for the Community.
DOESTHIS SOLVE ALL OF OUR CONS?
Pros
• Easily curatable content
• Securable
• Reliable
• Keeps DBA salaries high
Cons
• ASHX handlers can be slow
• Databases get gigantic and hugely
difficult to maintain
• Transformations take up valuable
server time
• Web server is busy handling media,
not other web requests
• File synchronization can be an
issue for web farm environments
• The web server exists in one place
geographically
• Not cloud friendly
SUGCON NORTH AMERICA 2015 31
32. Organized by the Community, for the Community.SUGCON NORTH AMERICA 2015 32
THANKYOU FOR COMING
Proof of concept code location
33. Organized by the Community, for the Community.SUGCON NORTH AMERICA 2015 33
QUESTIONS?