Matt Oswalt made two great points while tweeting about my Automation Gone Wild blog post:
However, life is not always rosy, so @stupidengineer asked:
Read more ...No one knows for sure how pervasive deep learning and artificial intelligence are in the aggregate across all of the datacenters in the world, but what we do know is that the use of these techniques is growing and could represent a big chunk of the processing that gets done every millisecond of every day.
We spend a lot of time thinking about such things, and as Nvidia was getting ready to launch its new Tesla P4 and P40 GPU accelerator cards, we asked Ian Buck, vice president of accelerated computing at Nvidia just how much computing could be devoted …
Nvidia Pushes Deep Learning Inference With New Pascal GPUs was written by Timothy Prickett Morgan at The Next Platform.
Stop using legacy mobile audio, especially for conference calls. There are better alternatives. You’re doing your customers and colleagues a disservice by using mobile audio. It’s time we moved on. PSTN is not much better either – switch to VoIP, and give your ears a break from crappy audio connections.
There are many different methods of encoding speech for transmission across networks. There are trade-offs with each, balancing bandwidth, voice quality, and endpoint requirements. The interesting point is that there is not a direct relationship between bandwidth and quality. Half the bandwidth does not have to mean half the quality.
The Mean opinion score test provides a way of ‘scoring’ the quality of a call. 1 is Bad, 5 is Excellent. G.711 encoding has a score of 4.1, which is very good quality, but uses 64kbps per call. GSM has a score of 3.5, which is the minimum acceptable level…but it only uses 12.2kbps. Pretty good tradeoff if you’re in a bandwidth-constrained environment.
But we’re no longer constrained by bandwidth. We don’t need to squeeze that audio call down to only a few kbps. We can use other options such as FaceTime, Continue reading
This post is a follow-up on my earlier post on using an SSH bastion host. Since that article was published, I’ve gotten some additional information that I wanted to be sure to share with my readers. It’s possible that this additional information may not affect you, but I’ll allow you to make that determination based on your use case and your specific environment.
You may recall that my original article said that you needed to enable agent forwarding, either via the -A command-line switch or via a ForwardAgent line in your SSH configuration file. This is unnecessary. (Thank you to several readers who contacted me about this issue.) I tested this several times using AWS instances, and was able to transparently connect to private instances (instances without a public IP address) via a bastion host without enabling agent forwarding. This is odd because almost every other tutorial I’ve seen or read instructs readers to enable agent forwarding. I’ve not yet determined why this is the case, but I’m going to do some additional testing and I’ll keep readers posted as I learn more.
Note that I’ve updated the original article accordingly.
The Continue reading
EMC paid $1.7B for Documentum, back when.
Jefferson County, Colorado (“Jeffco”) is a local jurisdiction located against the beautiful Rocky Mountains and adjacent to the state capital in Denver. Jeffco’s IT 
organization is charged with meeting the needs not only of the various internal departments of the county, but also of serving its half million residents.
As with most IT departments, Jeffco’s IT team has some key priorities to address, including modernizing application infrastructures and bringing more efficiency to business processes — all while fundamentally enhancing security. It was these needs that led Jeffco to VMware NSX. “We’re doing as much as we can to simplify our infrastructure, yet provide more security, higher up time, and better performance,” says Matt Alexander, Senior Systems Administrator.
Like many other organizations, Jeffco first considered VMware NSX for micro-segmentation. Their network had followed the traditional model of data center security: perimeter firewalls, DMZ, internal security zone. But this legacy security model wasn’t enough. Jeffco recognized the need to treat all network traffic — regardless of whether it originated inside or outside the data center — as potentially insecure. “From a micro-segmentation and east-west firewalling perspective, we may have had the ability [in the past] but it was exceptionally expensive with physical Continue reading
Microsoft has secured some big allies in a fight against the federal government, including three of its chief rivals, plus a hometown airline.
Microsoft is fighting the government over its right to tell customers when federal agents request their data and emails. The company filed a lawsuit in April against the federal government, charging such gag orders violate the Constitution and threaten the future of cloud computing.
+ Also on Network World: Microsoft says tech companies ‘whipsawed’ by conflicting laws on global data transfer +
To read this article in full or to leave a comment, please click here
The post Worth Reading: Hardware slaves to the master algorithm appeared first on 'net work.
As founding member of the LoRa Alliance, it's no surprise SoftBank is deploying a LoRaWAN network.
The HPE company now sells subscription-based networking products.
In the last post on this topic, we traced how snaproute’s BGP code moved to the open state. At the end of that post, the speaker encodes an open message using packet, _ := bgpOpenMsg.Encode(), and then sends it. What we should be expecting next is for an open message from the new peer to be received and processed. Receiving this open message will be an event, so what we’re going to need to look for is someplace in the code that processes the receipt of an open message. All the way back in the fifth post of this series, we actually unraveled this chain, and found this is the call chain we’re looking for—
I don’t want to retrace all those steps here, but the call to func (st *OpenSentState) processEvent() (around line 444 in fsm.go) looks correct. The call in question must be a call to a function that processes an event while the peer is in the open state. This call seems to satisfy both Continue reading
Verizon adds another IoT company to its portfolio.
We have heard about a great number of new architectures and approaches to scalable and efficient deep learning processing that sit outside of the standard CPU, GPU, and FPGA box and while each is different, many are leveraging a common element at all-important memory layer.
The Hybrid Memory Cube (HMC), which we expect to see much more of over the coming year and beyond, is at the heart of several custom architectures to suit the deep learning market. Nervana Systems, which was recently acquired by Intel (HMC maker, Micron’s close partner), Wave Computing, and other research efforts all see a …
Deep Learning Architectures Hinge on Hybrid Memory Cube was written by Nicole Hemsoth at The Next Platform.