Cloudflare Introduces the League of Entropy
Cloudflare, a business that powers Internet requests for more than 1 billion IP addresses per day, has introduced its solution for, what they say, is a long standing problem in cryptography. The demand for a constant generation of substantially large, unpredictable numbers, used in areas such as cryptography, money lotteries, and quantum mechanics, has increased substantially. Cloudflare’s solution? The League of Entropy.
The League is made up of a number of global organizations and individual contributors including: Cloudflare itself, Protocol Labs’ Nicolas Gailly, the University of Chile, and EPEI researchers Phillip Jovanovic and Ludovic Barman to name a few. In the past, single source randomness has been used to generate numbers, but it has the capability of being exploited to generate biased results. The League of Entropy is designed to hopefully thwart that, by creating a global network of servers that will, together, act as a beacon called dee-rand.
The beacon will be used to generate publicly verifiable random values every 60 seconds. Clients will also be able to use the drand beacon to request private randomness from some or all of the drand nodes which can reduce the chance of the encryption being cracked. Since the servers are on a global network, even if a few servers go offline, the beacon will continue to produce new numbers with the remaining online servers. This is only the beginning for the league. Cloudflare noted that it is open to others interested in joining their efforts in the future.
Google’s Solution to Keep Data Sets Private
After years of work, Google has come up with a technique to enable “multiparty computation” without exposing sensitive information to anyone that isn’t supposed to have access to it. The company will be releasing an open source cryptographic tool known as Private Join and Compute, which will facilitate the process of joining numeric columns from different data sets to calculate a sum, count, or average while keeping the data encrypted and unreadable during the process. Only the results can be decrypted and viewed by all parties.
Google has achieved this by repurposing and updating cryptographic principles from the 1970s and 90s. “Communicative encryption” allows the data sets to be encrypted with multiple keys without it mattering which order the keys are used, while “homomorphic encryption” enables a system to combine two encrypted data sets to determine their similarities and perform computations directly on the encrypted data.
The system is still considered to be computationally intensive, making it not feasible for use in some situations. That hasn’t stopped Google from spreading the buzz around their new technique. Since the product was announced, the company has been distributing a technical paper describing Private Join and Compute to academic and industry cryptographers.
Hyperledger Welcomes New Members
Hyperledger, an open source collaborative effort created to advance cross-industry blockchain technologies by the Linux Foundation, has announced the addition of eight new members including organizations such as: Microsoft, the Ethereum Foundation, Nomickel, and Salesforce. Hyperledger allows organizations to create solid, industry-specific apps, platforms, and hardware systems to support their individual business transactions by offering enterprise-grade, open-source distributed ledger frameworks and code bases.
Brian Behlendorf, Executive Director of Hyperledger stated, “The mix of blue-chip technology companies, international organizations, logistics and manufacturing players, and academia have joined us from a point around the world show the widespread interest in investment in open source enterprise blockchain technologies. This broad-based involvement is what drives our expanding portfolio of open source projects, which are fully conceived, developed and advanced by the ever-growing Hyperledger community.”
The collaborative group also announced they have added a new identity infrastructure project spotlighting interoperability, cross-project development and broad community involvement. Hyperledger will be gathering in Tokyo at the end of July for its annual members summit.
Descartes Labs System Sets New Benchmark
The HPC cluster uses over 40,000 processing cores and 158 terabytes of memory. It also touts more compute power than any system on the TOP500 list in 2010.
The company stated they didn’t ask Amazon to give their engineers any special dispensation, discount, or custom planning or setup. They simply wanted to see if they could complete it on their own. Which, if completed successfully, would also be a testament to the self-service model of AWS.
The potential for democratization of HPC was palpable since the cost to run custom hardware at that speed is probably closer to $20–30 million. Not to mention the 6–12 month wait time to build and validate the system.
In related news from Azure, the cloud provider announced general availability of its HC-series Virtual Machines in the West US 2 and East US regions. HC-series VMs are optimized for the most at-scale, computationally intensive HPC applications. Microsoft says it has demonstrated its HC-series VMs scaling to over 20,000 cores for a single application.
That’s a wrap for this week’s Liftr Cloud Look Ahead. Has your business made major strides using the cloud? We want to hear from you! Email us at ideas@liftrcloud.com.
See you next week!
