Amazon’s Cloud Computing Efforts
- 1. Amazon’s Cloud Computing Efforts: S3 and EC2 By: Chris Sosa
- 2. Materials Provided By “ An Evaluation of Amazon’s Grid Computing Services: EC2, S3 and SQS” by Garfinkel When I refer to tests, I mean from this paper Blogs Miscellaneous articles Amazon’s Web Services website
- 3. Motivation Producing a large accessible computing and storage infrastructure is expensive Physical Space, cooling, power, management Money Experience Large maintenance costs Never sure how much we actually need
- 4. Utility Computing Basically outsourcing infrastructure to the big guys It’s all about economy of scale Goal to make web-scale computing easier … how many buzz words does it take to fill a gallon jar?
- 5. Amazon’s Cloud Computing Overview Amazon’s Elastic Compute Cloud (EC2) Provides resizable / on-demand computing capacity Amazon’s Simple Storage Service (S3) Storage for the Internet Provides web services interface to store and retrieve any amount of data at any time and from anywhere (for a price)
- 6. Amazon’s Web Service Model Interface Web-based dashboard REST API – standard HTTP(/s) GET, PUT, DELETE commands Remote procedure calls through SOAP Variety of tools that perform basic operations Security Each request is signed 40-character secret key with HMAC X.509 certificates either given to or created by Amazon Signed timestamps to prevent replay attacks Can use TLS / SSL
- 7. First Up – EC2 Amazon uses Linux-based Virtual Machines running on Xen Actual resources are virtualized So much so that they are not even backed by real hardware “ Predictably provided” Two security mechanisms Write ssh public key into the VM image Firewall at the server side
- 8. EC2 Cost Model Instances $0.10 - Small Instance (per hour) 1.7 GB of memory, 1 EC2 Compute Unit (1 virtual core - 1.7 GHz processor), 160 GB of instance storage, 32-bit platform $0.40 - Large Instance (per hour) 7.5 GB of memory, 4 EC2 Compute Units (2 virtual cores with 2 EC2 Compute Units each), 850 GB of instance storage, 64-bit platform $0.80 - Extra Large Instance (per hour) 15 GB of memory, 8 EC2 Compute Units (4 virtual cores with 2 EC2 Compute Units each), 1690 GB of instance storage, 64-bit platform Data Transfer $0.10 per GB - all data transfer in $0.18 per GB - first 10 TB / month data transfer out $0.16 per GB - next 40 TB / month data transfer out $0.13 per GB - data transfer out / month over 50 TB
- 9. The Good Only pay for what you need “Guaranteed” to be 99.9% reliable (good enough?) (targets 99.99%) Fast, responsive and reliable Only takes a couple minutes to start a new instance Data transfer free between S3 and EC2
- 10. The Bad Still pay for computing you aren’t using but still running No persistent storage Can use S3 Writing to S3 slower and literally more expensive Poor termination clause Only given 60 day notice
- 11. Second – S3 250 Mb/second network link Objects stored in buckets Buckets have own namespace Up to 100 buckets per account Unlimited objects per bucket 5 GB limit on size of objects Computes MD5 checksum’s to check if transfers are correct Up to application to retry ACL’s by accounts / certificates allowed per object and per bucket (two-level) Basic interface to read/write/create/delete/chmod etc
- 12. S3 Pricing Storage $0.15 per GB-Month of storage used Data Transfer $0.10 per GB - all data transfer in $0.18 per GB - first 10 TB / month data transfer out $0.16 per GB - next 40 TB / month data transfer out $0.13 per GB - data transfer out / month over 50 TB Requests $0.01 per 1,000 PUT or LIST requests $0.01 per 10,000 GET and all other requests* * No charge for delete requests
- 13. The Pretty Available during 100% of tests Performance Works well for large file sizes Approaches maximum of 30 Mb/s with multiple threads (6) for test in paper Performs at about 10 MB /s with one thread doing one large data push
- 14. The Pretty Ugly No guarantee of persistence in data Will real companies buy into this? Claimed throughput only achievable with data transfers of 16MB or larger High transaction overhead Amazon has declined to explain why! Same Service Level Agreement terms as EC2
- 15. Conclusion EC2 and S3 provide applications with the ability to allocate computing and storage resources on demand This IS useful Cheaper than the alternative for reasonably sized companies Most claims of service have been validated Issues with bandwidth claims for S3 Issues with its Service Level Terms may plague EC2 and S3 from becoming more utilized
- 16. Questions