In the deep learning inferencing game, there are plenty of chipmakers, large and small, developing custom-built ASICs aimed at this application set. …
AI Chip Startup Puts Inference Cards on the Table was written by Michael Feldman at .
AI manages your money: Artificial Intelligence may eventually replace your financial advisor, Forbes suggests. AI can already spot financial trends really fast, but it may eventually compete with the personal touch of a human advisor, the story says. “Because artificial intelligence learns so much faster than humans, it is simply a matter of time before artificial intelligence can read human nuances and have an emotional intelligence quotient that exceeds those of most humans. When that happens, in the next few years, financial advisers will have a hard time competing based on personal relationships.”
Banning news: Russia has moved to ban what the government defines as fake news, joining several other countries headed in the same direction, the Boston Globe reports. A second law bans the publication of information showing disrespect to government bodies and officials. The fake news law allows fines of up to US$15,000.
Less fake, more news: Despite headlines about the spread of fake news during the 2016 U.S. elections, a majority of U.S. residents didn’t see fake news on social media, two recent studies suggest. On Twitter, fake news appeared on the feeds of just 1.1 percent of users, according to one study detailed in Continue reading
In its eighth-annual SIP survey, The SIP School finds that SIP trunking problems persist.
The crazy pace of webinar sessions continued last week. Howard Marks continued his deep dive into Hyper-Converged Infrastructure, this time focusing on go-to-market strategies, failure resiliency with replicas and local RAID, and the eternal debate (if you happen to be working for a certain $vendor) whether it’s better to run your HCI code in a VM and not in hypervisor kernel like your competitor does. He concluded with the description of what major players (VMware VSAN, Nutanix and HPE Simplivity) do.
On Thursday I started my Ansible 2.7 Updates saga, describing how network_cli plugin works, how they implemented generic CLI modules, how to use SSH keys or usernames and passwords for authentication (and how to make them secure), and how to execute commands on network devices (including an introduction into the gory details of parsing text outputs, JSON or XML).
The last thing I managed to cover was the cli_command module and how you can use it to execute any command on a network device… and then I ran out of time. We’ll continue with sample playbooks and network device configurations on February 12th.
You can get access to both webinars with Standard ipSpace.net subscription.
The case for network-accelerated query processing Lerner et al., CIDR’19
Datastores continue to advance on a number of fronts. Some of those that come to mind are adapting to faster networks (e.g. ‘FARM: Fast Remote Memory’) and persistent memory (see e.g. ‘Let’s talk about storage and recovery methods for non-volatile memory database systems’), deeply integrating approximate query processing (e.g. ‘ApproxHadoop: Bringing approximations to MapReduce frameworks’ and ‘BlinkDB’), embedding machine learning in the core of the system (e.g. ‘SageDB’), and offloading processing into the network (e.g KV-Direct) — one particular example of exploiting hardware accelerators. Today’s paper gives us an exciting look at the untapped potential for network-accelerated query processing. We’re going to need all that data structure synthesis and cost-model based exploration coupled with self-learning to unlock the potential that arises from all of these advances in tandem!
NetAccel uses programmable network devices to offload some query patterns for MPP databases into the switch.
Thus, for the first time, moving data through networking equipment can contributed to query execution. Our preliminary results show that we can improve response times on even the best Continue reading
This is a follow-up to my previous article, Arista BGP EVPN Overiew and Concepts. In the previous article, I discussed some terminologies and behavior of EVPN and the reason why EVPN is valuable in Data Center and Campus networks. Since then, I’ve learned how valuable it is in Service Provider networks as well, but I’ll […]
The post Arista BGP EVPN – Configuration Example appeared first on Overlaid.
Let’s talk about transmission overhead.
For various types of communications protocols, ranging from Ethernet to Fibre Channel to SATA to PCIe, there’s typically additional bits that are transmitted to help with error correction, error detection, and/or clock sync. These additional bits eat up some of the bandwidth, and is referred to generally as just “the overhead”.
For 1 Gigabit Ethernet and 8 Gigabit Fibre Channel as well as SATA I, II, and III, they use 8/10 overhead. Which means for every eight bits of data, an additional two bits are sent.
The difference is who pays for those extra bits. With Ethernet, Ethernet pays. With Fibre Channel and SATA, the user pays.
1 Gigabit Ethernet has a raw transmit rate of 1 gigabit per second. However, the actual transmission rate (baud, the rate at which raw 1s and 0s are transmitted) for Gigabit Ethernet is 1.25 gigabaud. This is to make up for the 8/10 overhead.
SATA and Fibre Channel, however, do not up the baud rate to accommodate for the 8/10 overhead. As such, even though 1,000 Gigabit / 8 bits per byte = 125 MB/s, Gigabit Fibre Channel only provides 100 MB/s. 25 MB/s is eaten up Continue reading
The public has until 31 January to comment on a draft set of rules in India that could result in big changes to online security and privacy.
The Indian government published the draft Information Technology [Intermediary Guidelines (Amendment) Rules] 2018, also known as the “Intermediary Rules” for public comment.
When it comes to the Internet, intermediaries are companies that mediate online communication and enable various forms of online expression.
The draft Intermediary Rules would change parts of the Information Technology Act, 2000 (the “IT Act”), which sets out the requirements intermediaries must meet to be shielded from liability for the activities of their users. The draft rules would also expand the requirements for all intermediaries, which are defined by the Indian government and include Internet service providers, cybercafés, online companies, social media platforms, and others. For example, all intermediaries would have to regularly notify users on content they shouldn’t share; make unlawful content traceable; and deploy automated tools to identify and disable unlawful information or content, among other new requirements.
Here’s some more background:
As champions of an open, globally-secure, and trusted Internet, International Data Privacy Day is a big deal around these parts.
But making sure you’re able to share what you want, when you want, should be something the world stands for more than once a year. Every day should be Data Privacy Day.
These days, it feels all too common to hear stories about policy or law enforcement officials trying to create backdoors into technologies like encryption. These backdoors could put our online security at risk.
Just a little over one month ago, Business Insider reported that smart home devices dominated Christmas 2018 sales on Amazon, while the Alexa app, which enables people to control those smart devices, was the most downloaded on Google Play and the Apple App store on Christmas Day.
As the Internet becomes more and more a part of our everyday lives, each of us can take actions to ensure that privacy and security are a top priority.
Let’s come together on Data Privacy Day to celebrate the possibilities an open, globally connected, trusted, and secure Internet brings. Here are ways you can help make it happen where you live:
(And don’t forget to make a cake! Continue reading
Data Privacy Day is a little like celebrating the anniversary of your first date.
They are both a yearly occasion to reflect on the most important relationships in our life, the former with those who know the most about us, the latter with our significant other.
It’s also a reminder that relationships are built on trust – and how fragile that trust can be.
Privacy online relies on trust at its core. But as we become more reliant on connected devices and virtual assistants to handle our most intimate health, banking, and private information, we’re putting our trust into shaky hands.
Honesty is the foundation of trust and it’s just as important in our relationships with loved ones as those with data brokers. It’s crucial for data brokers to be honest with users about who, when, and how people have access to their personal data, especially as we transition into smarter homes and cities.
Let’s face it: there’s a huge market for the information we share online. Both U.S. and Canadian Internet companies are increasingly trying to collect our personal data – whether we know it or not.
It’s clear we want more control over our privacy, but each Continue reading
One of the most common mistakes/confusion I see with regard to storage is how speed is measured.
In tech, there’s some cultural conventions to which units speeds are discussed in.
Of course they both say the same thing, just in different units. You could measure bytes per second in the networking world and bits per second in the server/storage world, but it’s not the “native” method and could add to confusion.
For NAS, we have a bit of a conundrum in that we’re talking about both worlds. So it’s important to communicate effectively which method you’re using to measure speed: bits of bytes.
Generally speaking, if you want to talk about Bytes, you capitalize the B. If you want to talk about bits, the b is lower case. I.e. 100 MB/s (100 Megabytes per second) and 100 Mbit or Mb (100 Megabit per second).
This is important, because there a 8 bits in a byte, the difference in speed is pretty stark depending on if you’re talking about bits per second or bytes per Continue reading
I maintain two git repositories with various network and vendor icons in SVG format. Feel free to use them, but do request the vendors permission to use their trademark images, if what you are using them for is something that requires permission. Network Icons Sample Vendor Icons...continue reading
Computing power and big data are fundamental to the bioinformatic research being carried out by the Leadership Computing Facility at the Oak Ridge National Laboratory’s (ORNL) in Tennessee. …
ORNL’s Summit Supercomputer Targets Opioid Addiction was written by Nicole Hemsoth at .
Network Engineers create and operate prototype networks all the time. Prototype networks are used to validate designs, test features or changes, troubleshoot use-case scenarios, and often just for learning. Typically, pre-prod testing environments are set up in such a way that device host names, attributes, configurations, IP assignments, software versions, and topologies are mostly inconsistent with production environments. This inconsistency is counter-intuitive, considering that accurate design validations should closely match reality to avoid any mistakes when deploying in production.
Cumulus Linux can run as a virtual appliance, allowing network engineers to build to-scale virtual networks for activities like modeling changes and performing validations, while opening the door for similar DevOps methodologies application developers have operated with for years: validated testing before deploying in production for continuous integration.
Cumulus VX (Virtual Experience) is a Cumulus Linux virtual appliance. You can test drive Cumulus Linux on a laptop, while those fluent with Cumulus Linux can prototype large networks and develop software integrations before deploying into production environments.
Cumulus VX is a platform — just like Cumulus Linux on a real switch — and therefore is designed to perform just like an actual switch running Cumulus Linux. Every feature you Continue reading