I really dislike corporate VPNs that don’t allow split tunneling—disconnecting from the VPN to print on a local printer, or access a local network attached drive, puts a real crimp in productivity. In the case of services reachable over both IPv6 and IPv4, particularly if the IPv6 path is preferred, split tunneling can be quite dangerous, as explained in RFC7359. Let’s use the network below to illustrate.
In this network, host A is communicating with server B through a VPN, terminated by the VPN concentrator marked as “VPN.” Assume the host is reachable on both 192.0.2.1 and 2001:fb8:0:1::1. The host, the upstream router, the network in the cloud, and the server are all IPv6 reachable. When the host first connects, it will attempt both the IPv6 and IPv4 connections, and choose to use the IPv6 connection (this is what most current operating systems will do).
The problem is: the VPN connection doesn’t support IPv6 at all—it only supports IPv4. Because IPv6 is preferred, the traffic between the host and the server will take the local IPv6 connection, which is not encrypted—the blue dash/dot line—rather than the encrypted IPv4 tunnel—the red dashed line. The user, host, and Continue reading
Hotjar offers free website analytics so they have a challenging mission: handle hundreds of millions of requests per day from mostly free users. Marc von Brockdorff, Co-Founder & Director of Engineering at Hotjar, summarized the lessons they've learned in: 9 Lessons Learned Scaling Hotjar's Tech Architecture To Handle 21,875,000 Requests Per Hour.
In response to the criticism their architecture looks like a hot mess, Erik Näslund, Chief Architect at Hotjar, gives the highlights of their architecture:
The post Worth Reading: The great DevOps train wreck appeared first on 'net work.
Docker containers are clearly transforming nearly every aspect of IT these days – and networking is no exception.
DockerCon 2016 was packed lots with great conference sessions! Attendees enjoyed the variety of topics in the agenda including advanced technical deep dives in the Black Belt track and practical applications of Docker in the Use Case track.
Through the mobile app (powered by Docker of course!), DockerCon 2016 attendees voted on their favorite sessions at the conference. Here are 5 of the top 10 highest rated sessions at DockerCon 2016: Continue reading
In the last couple of years, we have written and heard about the usefulness of GPUs for deep learning training as well as, to a lesser extent, custom ASICs and FPGAs. All of these options have shown performance or efficiency advantages over commodity CPU-only approaches, but programming for all of these is often a challenge.
Programmability hurdles aside, deep learning training on accelerators is standard, but is often limited to a single choice—GPUs or, to a far lesser extent, FPGAs. Now, a research team from the University of California Santa Barbara has proposed a new middleware platform that can combine …
Emerging “Universal” FPGA, GPU Platform for Deep Learning was written by Nicole Hemsoth at The Next Platform.
When writing an HTTP server or client in Go, timeouts are amongst the easiest and most subtle things to get wrong: there’s many to choose from, and a mistake can have no consequences for a long time, until the network glitches and the process hangs.
HTTP is a complex multi-stage protocol, so there's no one-size fits all solution to timeouts. Think about a streaming endpoint versus a JSON API versus a Comet endpoint. Indeed, the defaults are often not what you want.
In this post I’ll take apart the various stages you might need to apply a timeout to, and look at the different ways to do it, on both the Server and the Client side.
First, you need to know about the network primitive that Go exposes to implement timeouts: Deadlines.
Exposed by net.Conn
with the Set[Read|Write]Deadline(time.Time)
methods, Deadlines are an absolute time which when reached makes all I/O operations fail with a timeout error.
Deadlines are not timeouts. Once set they stay in force forever (or until the next call to SetDeadline
), no matter if and how the connection is used in the meantime. So to build a timeout with SetDeadline
you'll have to Continue reading