Issue link: https://read.dmtmag.com/i/43993
Data transfer into & out of cloud Data transfer into & out of cloud Data transfer into & out of cloud Storage 30 days Processing 30 days Storage 30 days Processing 30 days Storage 30 days Processing 30 days Data transfer into & out of cloud Data transfer into & out of cloud Data transfer into & out of cloud Storage 30 days Processing 1 day Storage 30 days Processing 1 day Storage 30 days Processing 1 day Data transfer into & out of cloud Data transfer into & out of cloud Data transfer into & out of cloud Storage 1 day Processing 1 day Storage 1 day Processing 1 day Storage 1 day Processing 1 day lFigure 1. For three cloud-computing scenarios, the inner circle assumes Low-Cost pricing, and the outer circle assumes High- Cost pricing. In Scenario 1 (top), the data are processed in the cloud for 30 days and stored in the cloud for 30 days. Low-Cost pricing is $310.80; High-Cost pricing is $1,692. In Scenario 2 (middle), the data are processed in the cloud for one day and stored in the cloud for 30 days. Low-Cost pricing is $300.36; High-Cost pricing is $578.40. In Scenario 3 (bottom), the data are processed in the cloud in one day and stored in the cloud for one day. Low-Cost pricing is $184.36; High-Cost pricing is $365.73. (Badger et al) 16 GEO W ORLD / SEPTEMBE R 2O11 is $100, for a total of $180 in transfer costs. Under the assumption of High Cost, transferring 1TB into the cloud is $100 and out of the cloud is $220, for a total of $320. Those constant dollars become a larger component of total cost as storage and processing are reduced. For organizations with a good understanding of the current total cost of ownership for on-premise data centers, the "Amazon EC2 Cost Comparison Calculator" from Amazon Web Services (AWS) can be used to model costs on the AWS infrastructure (awsmedia.s3.amazonaws.com/Amazon_EC2_Cost_ Comparison_Calculator_042810.xls). For large datasets, transferring data to "the cloud" is best done by shipping a physical storage device to the vendor. A file size of 100MB takes nearly 10 minutes on a T1 line, while 1GB takes more than 1.5 hours over the same line. As faster Internet connections become more widely available, the file size that can reasonably be transmit- ted will increase. But, for the foreseeable future, sen- sor resolution will continue to improve at such a rate that typical file sizes for geospatial imagery still will be most efficiently transferred by shipping a device. Barriers and Benefits James Fee, chief evangelist for WeoGeo, notes that the mindset of information technology (IT) professionals has been "I need a server. I can see the server. I control that server. I'm doing my job." So thinking "My life will be easier if I let others worry about choosing the hardware and dealing with power outages and protecting against hackers" represents a significant shift in thinking. In a similar vein, Charles Babcock, in his book, Management Strategies for the Cloud Revolution, identi- fies the No. 1 barrier to adoption as the opposition of CEOs to the name "cloud computing." Babcock says the name itself ranks ahead of concerns about data security and the privacy of user identities. Fee notes that the phrase "hosted services" is more informative than "cloud computing," so he uses that to start conversations about this revolutionary capability. Fee identifies two major issues for people consider- ing cloud computing: 1) security and 2) lack of speed. WeoGeo recommends that customers with sensitive data don't put them in a public cloud. Also, WeoGeo's offerings are based on AWS, and Fee describes the read/write speed to Amazon as "like using a floppy." Industry Trends