Managing terabytes on aws

81

description

JFrog is handling huge amount of binaries files for all our customers. Since each customer has its own space and domain, using a global Object Store can be tricky. Also our application cannot work with an “eventually consistent” storage, and cannot deliver customer requirements with current S3 performance. Learn in this session how we managed fast upload, critical replication and backups, and global download availability of the terabytes of JFrog customer binaries files.

Transcript of Managing terabytes on aws

Page 1: Managing terabytes on aws
Page 2: Managing terabytes on aws
Page 3: Managing terabytes on aws
Page 4: Managing terabytes on aws
Page 5: Managing terabytes on aws
Page 6: Managing terabytes on aws
Page 7: Managing terabytes on aws
Page 8: Managing terabytes on aws
Page 9: Managing terabytes on aws
Page 10: Managing terabytes on aws
Page 11: Managing terabytes on aws
Page 12: Managing terabytes on aws
Page 13: Managing terabytes on aws

Request

Storage File Listing Checksum Storage

Page 14: Managing terabytes on aws

Request

Storage File Listing Checksum Storage

Metadata requests

Cache check, FS request

DB request

Page 15: Managing terabytes on aws

Request

Storage File Listing Checksum Storage

Metadata requests

Cache check, FS request

DB request

HEAD requests Cache check, FS request

DB request

Page 16: Managing terabytes on aws

Request

Storage File Listing Checksum Storage

Metadata requests

Cache check, FS request

DB request

HEAD requests Cache check, FS request

DB request

GET request Cache check, FS request

DB + FS request

Page 17: Managing terabytes on aws

Request

Storage File Listing Checksum Storage

Metadata requests

Cache check, FS request

DB request

HEAD requests Cache check, FS request

DB request

GET request Cache check, FS request

DB + FS request

PUT request FS request DB + FS request

Page 18: Managing terabytes on aws
Page 19: Managing terabytes on aws
Page 20: Managing terabytes on aws
Page 21: Managing terabytes on aws
Page 22: Managing terabytes on aws
Page 23: Managing terabytes on aws
Page 24: Managing terabytes on aws
Page 25: Managing terabytes on aws
Page 26: Managing terabytes on aws
Page 27: Managing terabytes on aws
Page 28: Managing terabytes on aws
Page 29: Managing terabytes on aws
Page 30: Managing terabytes on aws
Page 31: Managing terabytes on aws
Page 32: Managing terabytes on aws
Page 33: Managing terabytes on aws
Page 34: Managing terabytes on aws
Page 35: Managing terabytes on aws
Page 36: Managing terabytes on aws
Page 37: Managing terabytes on aws
Page 38: Managing terabytes on aws

EBS + LVM

= WIN!

Page 39: Managing terabytes on aws

/data

250GB 250GB

Page 40: Managing terabytes on aws

/data

250GB 250GB 250GB 250GB

Page 41: Managing terabytes on aws

/data

250GB 250GB 250GB 250GB 250GB 250GB

Page 42: Managing terabytes on aws
Page 43: Managing terabytes on aws
Page 44: Managing terabytes on aws
Page 45: Managing terabytes on aws
Page 46: Managing terabytes on aws
Page 47: Managing terabytes on aws
Page 48: Managing terabytes on aws
Page 49: Managing terabytes on aws
Page 50: Managing terabytes on aws
Page 51: Managing terabytes on aws

Backup files

Page 52: Managing terabytes on aws

Backup files

Backup database

Page 53: Managing terabytes on aws

Backup files

Backup database

Keep them in sync!

Page 54: Managing terabytes on aws

Backup files

Backup database

Keep them in sync!

Page 55: Managing terabytes on aws

Backup only deltas

Page 56: Managing terabytes on aws

Backup only deltas

Manage threads to

ec2 efficiently

Page 57: Managing terabytes on aws

Backup only deltas

Manage threads to

ec2 efficiently

Keep the data

isolated

Page 58: Managing terabytes on aws

Backup only deltas

Manage threads to

ec2 efficiently

Keep the data

isolated

Page 59: Managing terabytes on aws
Page 60: Managing terabytes on aws
Page 61: Managing terabytes on aws
Page 62: Managing terabytes on aws
Page 63: Managing terabytes on aws
Page 64: Managing terabytes on aws
Page 65: Managing terabytes on aws

Customer EBS

Customer EBS

Customer EBS

Page 66: Managing terabytes on aws

De-duped

EBS

HA Storage

Customer EBS

Customer EBS

Customer EBS

Page 67: Managing terabytes on aws
Page 68: Managing terabytes on aws

De-duped Glacier

De-duped

EBS Backup

HA Storage

Customer EBS

Customer EBS

Customer EBS

Page 69: Managing terabytes on aws
Page 70: Managing terabytes on aws
Page 71: Managing terabytes on aws
Page 72: Managing terabytes on aws
Page 73: Managing terabytes on aws
Page 74: Managing terabytes on aws

Create empty Artifactory instance

Page 75: Managing terabytes on aws

Create empty Artifactory instance

Populate from back

storage on demand

Page 76: Managing terabytes on aws

Create empty Artifactory instance

Populate from back

storage on demand

Drink

Cosmopolitan

Page 77: Managing terabytes on aws

Conclusions time

1. Be lazy.

Page 78: Managing terabytes on aws
Page 79: Managing terabytes on aws

Conclusions time

1. Be lazy.

2. Be smart.

Page 80: Managing terabytes on aws

Conclusions time

1. Be lazy.

2. Be smart.

3. Be agile.

Page 81: Managing terabytes on aws

Drink

Cosmopolitan