Language Selection

English French German Italian Portuguese Spanish

BitTorrent Creator Dismisses M$ P2P Project

Filed under
Microsoft
Software

Yesterday, BitTorrent creator Bram Cohen called Microsoft Research's attempt to create its own version of the person-to-person application "vaporware" and "complete garbage."

In Cohen's blog, he said Microsoft Corp.'s boast that the new P2P protocol, codenamed Avalanche, would fix transfer rate problems and disconnections was unfounded.

Cohen added that since the Microsoft experiments were done via "simulations," the results were flawed-the tests do not take varying transfer rates into consideration, nor the limitations of all the users' computers.

"Particularly worrisome for their proposed scheme is disk access," Bram said in his blog. "If the size of the file being transferred is greater than the size of memory, their entire system could easily get bogged down doing disk seeks and reads, since it needs to do constant recombinations of the entire file to build the pieces to be sent over the wire."

Despite the criticism, Avalanche still seems to be holding its own in the word-of-mouth department. While the Microsoft P2P protocol has a few different elements than BitTorrent, the premise remains the same: If someone needs to transfer a huge file, such as a video or a form of software, to many people, a server usually cannot handle the workload. Instead, the file "swarms" across the Web, and can be downloaded in bits and pieces from many destinations, each containing small pieces of the entire file.

The main problem with this system is that downloading a file can take a long time, because at the end of the process, users usually have to wait to find and download a last bit of information, called the "rarest bit." Sometimes waiting for the rarest bit can take 12 hours or more, depending on the popularity and size of the downloading file in question.

Microsoft, much to Cohen's disapproval, says it's found a way to avoid the waiting game by recoding all of the pieces of a file so that each one shared is a linear combination of the pieces. After a user has downloaded a few of these, the user can generate new combinations from the pieces and send those on to other peers.

At that point, instead of having to wait for more important pieces of the puzzle, any piece can be used to complete the entire picture. And, since the same information will no longer have to travel back and forth, overall network traffic will decrease as well. In fact, Avalanche researchers said this new transfer method made download times 20 to 30 percent faster.

Keeping in mind that this research is just that-research-David Card, an analyst at JupiterResearch of Jupitermedia Corp., said it's hard not to wonder why Microsoft would want to create its own P2P services, especially as Microsoft is actively involved in regaining and maintaining the trust of entertainment companies by taking an active role in digital rights management.

Full Story.

More in Tux Machines

Security Leftovers

  • efail: Outdated Crypto Standards are to blame
    I have a lot of thoughts about the recently published efail vulnerability, so I thought I'd start to writeup some of them. I'd like to skip all the public outrage about the disclosure process for now, as I mainly wanted to get into the technical issues, explain what I think went wrong and how things can become more secure in the future. I read lots of wrong statements that "it's only the mail clients" and the underlying crypto standards are fine, so I'll start by explaining why I believe the OpenPGP and S/MIME standards are broken and why we still see these kinds of bugs in 2018. I plan to do a second writeup that will be titled "efail: HTML mails are to blame". I assume most will have heard of efail by now, but the quick version is this: By combining a weakness in cryptographic modes along with HTML emails a team of researchers was able to figure out a variety of ways in which mail clients can be tricked into exfiltrating the content of encrypted e-mails. Not all of the attack scenarios involve crypto, but those that do exploit a property of encryption modes that is called malleability. It means that under certain circumstances you can do controlled changes of the content of an encrypted message. [...] Properly using authenticated encryption modes can prevent a lot of problems. It's been a known issue in OpenPGP, but until know it wasn't pressing enough to fix it. The good news is that with minor modifications OpenPGP can still be used safely. And having a future OpenPGP standard with proper authenticated encryption is definitely possible. For S/MIME the situation is much more dire and it's probably best to just give up on it. It was never a good idea in the first place to have competing standards for e-mail encryption. For other crypto protocols there's a lesson to be learned as well: Stop using unauthenticated encryption modes. If anything efail should make that abundantly clear.
  • Comcast Leaked Customer Wi-Fi Logins in Plaintext, Change Your Passcode Now
    A Comcast Xfinity website was leaking Wi-Fi names and passwords, meaning now is a good time to change your Wi-Fi passcode. The site, intended to help new customers set up new routers, could easily be fooled into revealing the location of and password for any customer’s Wi-Fi network. A customer ID and a house or apartment number was all would-be attackers needed to get full access to your network, along with your full address.
  • Update Fedora Linux using terminal for latest software patches
  • Patch for New Spectre-Like CPU Bug Could Affect Your Performance
  • container_t versus svirt_lxc_net_t

today's howtos

Red Hat News

  • “Ultimate Private Cloud” Demo, Under The Hood!
    At the recent Red Hat Summit in San Francisco, and more recently the OpenStack Summit in Vancouver, the OpenStack engineering team worked on some interesting demos for the keynote talks. I’ve been directly involved with the deployment of Red Hat OpenShift Platform on bare metal using the Red Hat OpenStack Platform director deployment/management tool, integrated with openshift-ansible. I’ll give some details of this demo, the upstream TripleO features related to this work, and insight around the potential use-cases.
  • Discover the possibilities of hybrid cloud during a joint virtual event with Red Hat & Microsoft [Ed: [Ed: When Red Hat pus Microsoft executives at top positions inside Red Hat...]
  • Red Hat OpenStack Customer Survey 2018: containers, technical support top of mind
    In 2016, we surveyed our customer base on their use of OpenStack in production, getting a pulse-check on the top considerations, expectations, and benefits of a Red Hat OpenStack Platform deployment. With 2018 marking five years of Red Hat OpenStack Platform, we checked back in with our customers to see if their experiences or expectations of OpenStack have changed. Our survey found:
  • Red Hat CEO Jim Whitehurst On How He Plans To Win The Container Market
  • Juniper, Red Hat Tighten Integration to Fend Off VMware
    Juniper Networks and Red Hat have tightened their integration efforts in a move to help ease enterprise adoption of cloud-native platforms and bolster their own offerings against the likes of VMware and Cisco. The latest platform integration includes the Red Hat OpenStack Platform; Red Hat’s OpenShift Container Platform running as a platform-as-a-service (PaaS) on top of or next to the OpenStack platform depending on deployment architecture; and Juniper’s Contrail Enterprise Multi-Cloud platform running as the networking and security layer to unify those together. This integration is designed as a managed system to help deploy and run applications and services on any virtual machine (VM), container platform, and any cloud environment.
  • Red Hat OpenStack HCI Targets Telco Hybrid Cloud, 5G Deployments
    Red Hat today rolled out a hyperconverged infrastructure (HCI) platform based on OpenStack compute and Ceph storage. The new product targets service providers looking to deploy virtual network functions (VNFs) and 5G technologies on top of open source software. Launched at this week’s OpenStack Summit, the Red Hat Hyperconverged Infrastructure for Cloud combines Red Hat OpenStack Platform 13 and Red Hat Ceph Storage 3 into one product. Red Hat says it is the largest contributor to both open source projects.
  • Red Hat Hyperconverged Infrastructure for Cloud Bridges Datacenters and Edge Deployments
  • GSoC 2018: Week 1
    This time, I am working on improving the Fedora Community App with the Fedora project. It’s been a week since we started off our coding on may 14. The Fedora App is a central location for Fedora users and innovators to stay updated on The Fedora Project. News updates, social posts, Ask Fedora, as well as articles from Fedora Magazine are all held under this app.

Today in Techrights