Loading...
Please wait, while we are loading the content...
Similar Documents
On the ( Im ) Practicality of Securing Untrusted Computing Clouds with Cryptography
| Content Provider | Semantic Scholar |
|---|---|
| Author | Chen, Yao |
| Copyright Year | 2010 |
| Abstract | In a recent interview, Whitfield Diffie argued that “the whole point of cloud computing is economy” and while it is possible in principle for “computation to be d one on encrypted data, [...] current techniques would more than undo the economy gained by the outsour cing and show little sign of becoming practical”. In this paper we aim to understand whether this is truly the case and quantify just how expensive it is to secure data processing in untrusted, pote ntially curious clouds. We start by looking at the economics of computing in general a nd clouds in particular. Specifically, we derive the end-to-end cost of a CPU cycle in various enviro ments and show that its cost lies between 0.58 picocents in efficient clouds and 26.02 picocents for mall business deployment scenarios (1 picocent = $1 × 10), valuesvalidated against current cloud pricing. We then evaluate the cost of networking and show that, in order to offset the costs of netw orking, cloud computing makes economical sense only for compute intensive applications requiring at le st 3800 compute cycles per each 32 bits of transferred input. Finally, we explore the cost of common cryptography primiti ves as well as the viability of their deployment for cloud security purposes. We conclude that Di ffie was correct. Securing outsourced data and computation against untrusted clouds is indeed costlie r than the associated savings, with outsourcing mechanisms up to 5+ orders of magnitudes costlier than their non-outsourced locally run alternatives. This is simply because today’s cryptography does not allow f r e ficientoblivious processing of complex enoughfunctions on encrypted data. And outsourcing simple operat ions – such as existing research in querying encrypted data, keyword searches, sel ections, projections, and simple aggregates – is simply not profitable (too few compute cycles / input word to offset the client’s distance from the cloud). Thus, while traditional security mechanisms allow the elegant handling of inter-client and outside adversaries, today it is still too costly to secure against c loud insiders with cryptography. |
| File Format | PDF HTM / HTML |
| Alternate Webpage(s) | https://zxr.io/research/pcost2010-draft.pdf |
| Language | English |
| Access Restriction | Open |
| Content Type | Text |
| Resource Type | Article |
