NetworkingNexus.net

The Hard Limits for Deep Learning in HPC

If the hype is to be believed, there is no computational problem that cannot be tackled faster and better by artificial intelligence. But many of the supercomputing sites of the world beg to differ.

With that said, the deep learning boom has benefitted HPC in numerous ways, including bringing new cred to the years of hardware engineering around GPUs, software scalability tooling for complex parallel codes, and other feats of efficient performance at scale. And there are indeed areas of high performance computing that stand to benefit from integration of deep learning into the larger workflow including weather, cosmology, molecular …

The Hard Limits for Deep Learning in HPC was written by Nicole Hemsoth at The Next Platform.

Tech Talk: The beginning of the end of the password in 2018

Hackers can crack them, and users are careless with them. Passwords have long been problematic, but they are getting renewed attention after a series of major breaches. Our panelists look at the coming technology to help with password security.

Tech Talk: The coding benefits of serverless computing

The panel talks about serverless systems, which spin up a snippet of code that runs on demand to perform a business operation. It's a step toward a more developer-friendly approach to code development.

Tech Talk ep 5 pt 1: The beginning of the end of the password in 2018

tktktk

Apstra Adds Intent-Based Analytics to Its Operating System

It uses the same device agents for its intent-based networking.

phpIPAM on nginx

phpIPAM on nginx webserver

Using Docker Machine with Azure

I’ve written about using Docker Machine with a number of different providers, such as with AWS, with OpenStack, and even with a local KVM/Libvirt daemon. In this post, I’ll expand that series to show using Docker Machine with Azure. (This is a follow-up to my earlier post on experimenting with Azure.)

As with most of the other Docker Machine providers, using Docker Machine with Azure is reasonably straightforward. Run docker-machine create -d azure --help to get an idea of some of the parameters you can use when creating VMs on Azure using Docker Machine. A full list of the various parameters and options for the Azure drive is also available.

The only required parameter is --azure-subscription-id, which specifies your Azure subscription ID. If you don’t know this, or want to obtain it programmatically, you can use this Azure CLI command:

az account show --query "id" -o tsv

If you have more than one subscription, you’ll probably need to modify this command to filter it down to the specific subscription you want to use.

Additional parameters that you can supply include (but aren’t limited to):

Use the --azure-image parameter to specify the VM image you’d like to Continue reading

Automatic Product Pitch Generator

Because we all like a little bit of fun, I created an automatic Product Pitch Generator for network vendors. More accurately, a conversation about buzzwords occurred in the context of Networking Field Day 17 where it was jokingly suggested that we needed to take some of the amazing words we were hearing and make a generator. Here’s the best part (if you look at it this way); we ended up with two generators!

Big props to Jordan Martin who hacked together some Python, and created his Network Product Buzzword Generator which I think is hilarious (go try it out!). Meanwhile, I was hacking together some Go and came up with this mess of a Product Pitch Generator:

Hit Refresh to get a new Pitch!

And finally, if you’re watching the NFD17 livestream, why not ML-wash yourself and play NFD17 Buzzword Bingo?

Have fun!

If you liked this post, please do click through to the source at Automatic Product Pitch Generator and give me a share/like. Thank you!

Verizon Adds Virtual Session Border Controller to VNS Platform

It's based on Ribbon Communications' SBC software.

Addressing 2017

Time for another annual roundup from the world of IP addresses. Let’s see what has changed in the past 12 months in addressing the Internet and look at how IP address allocation information can inform us of the changing nature of the network itself.

Data could one day be stored on molecules

Billions of terabytes of data could be stored in one small flask of liquid, a group of scientists believe. The team from Brown University says soon it will be able to figure out a chemical-derived way of storing and manipulating mass-data by loading it onto molecules and then dissolving the molecules into liquids.If the method is successful, large-scale, synthetic molecule storage in liquids could one day replace hard drives. It would be a case of the traditional engineering that we’ve always pursued for storage being replaced by chemistry in our machines and data centers.Also on Network World: The future of storage: Pure Storage CEO Charlie Giancarlo shares his predictions The U.S. Department of Defense’s Defense Advanced Research Projects Agency (DARPA) has awarded the Brown team $4.1 million to work out how move the concept forward.To read this article in full, please click here

Data could one day be stored on molecules

Worth Reading: Quantum Computing and Cryptography

All encryption is based on substitution, from simple schemes we did as kids to complex non-repeating patterns based on a long numerical key. Modern encryption systems construct these keys using a combination of private and public keys with mathematical one-way functions. —Jonathan Homa @ECI

Limited Time Only: Read our Springer/Nature Paper on Healthcare, Security, and Privacy

Last year, I was invited to contribute a paper to a special edition of the Health and Technology Journal published by Springer/Nature. The special issue addressed privacy and security, with a particular focus on healthcare and medical data. I’m happy to announce that now, for four weeks only, the publishers have made the whole issue available free.

From our accompanying blog post last July:

“The paper, “Trust and ethical data handling in the healthcare context” examines the issues associated with healthcare data in terms of ethics, privacy, and trust, and makes recommendations about what we, as individuals, should ask for and expect from the organisations we entrust with our most sensitive personal data.”

Although we can find several comprehensive and mature data protection frameworks around the world, current legal safeguards to not seem to prevent data controllers from indulging in:

over collection
insufficient care of personal data
unexpected or unwelcome use
excessive sharing

In my paper, I argue that a narrow focus on regulatory compliance can lead to a “checklist” mentality, obscure the real reasons why organisations should treat data with care and respect, and lead to poor outcomes for both the organisation and the individual. I Continue reading

Getting started with Linux: the basics – part 3

In part 1 of our series on the basics of Linux, we learned how to download Linux, whether you should use the CLI or the GUI, how to get a SSH client, how to login to Linux, and how to get help.

In part 2 of our series, we learned how to know what type of Linux you are using and how to navigate the Linux file system.

In this final post in the series, you’ll learn about Linux files, Linux permissions, Linux package management, and how to install applications in Linux.

Files and permissions

Let’s say that a user named “david” was denied access to the file /var/log/syslog. The reason for that is likely the user “david” doesn’t have permission to access to the file.

You can see this if you execute ls -l /var/log/syslog:

david@debian:~$ ls -l /var/log/syslog -rw-r----- 1 root adm 9074 May 15 10:17 /var/log/syslog

The file is owned by the user “root” and the group “adm”. The file permissions are “rw” (shorthand for read/write) for the owner and “r” (shorthand for “read”) for the group with no permissions for anyone else. The graphic below shows how file permissions work in Linux.

Linux package management