deadgrey19

15 ms·

91.
by deadgrey19 12y ago
With some luck, it will be published early in the new year.
92.
by deadgrey19 12y ago
All good points. I can't say what the future of Naiad is. What they have done to Microsoft Research Silicon Valley is disgusting (I worked there too for a short time). In our experiences the performance claims with Spark have been more
93.
by deadgrey19 12y ago
As part of our work we have done extensive comparisons of Spark on various workloads, clusters and cluster sizes comparing with Hadoop Map Reduce, Naiad and several other frameworks. We've found Spark to be temperamental, hard to confi
94.
by deadgrey19 12y ago
"providing fingerprints and other biometric information is considered outside the protection" Is it just me, or is there a contradiction here? I'm happy to provide you with a finger print (in ink), but that in itself is not e
95.
by deadgrey19 12y ago
I wonder how it compares with seL4 ( http://sel4.systems/ ) which is verified, open source, "bare metal" and designed for minimality, security and performance.
96.
by deadgrey19 12y ago
Ethernet now also supports RDMA via the RoCE standard. No idea how it compares to Inifiband RDMA, but it has been heavily exploited in this project from Microsoft Research: https://www.usenix.org/system/files/confe
97.
by deadgrey19 12y ago
Thanks. Yes, I used to work at Zomojo (Zomojo.com), a prop-trading HFT firm which is also Exablaze's parent company. I'm now a doctoral researcher at the university of Cambridge although I do still maintain a good relationship wit
98.
by deadgrey19 12y ago
I think this is what they mean when they call it a "layer2+" device. 110ns is the device in its capacity as a full layer 2(+) switch. The "warp span" equivalent would be using the layer broadcast groups functionality wh
99.
by deadgrey19 12y ago
It's not just an FPGA, the 2x application bays support each support 48ports of connectivity to anything, at this point it's "just" a 48port FPGA implementing the worlds fastest layer 2+ switch, but it could be anything i
100.
by deadgrey19 12y ago
Yes. There's a layer one patch panel inside, with 48x3 ports. 48 for the front and 48x2 for inside. Inside they have two "application" bays and in one of them is a layer2+ switching device. So it's a layer 1+2+ device in
101.
by deadgrey19 12y ago
Agreed. Independent verification is definitely needed. Although I think that this is why they've announced this at STAC which does this ( https://stacresearch.com/about ). I've moved over to academia now, but I cam
102.
by deadgrey19 12y ago
I don't think Arista sells that device anymore. Having said that, the Arista device (7124FX?) only had the FPGA on 8 out of 24 ports and the latency through the transceivers was pretty terrible. Very few people took it up. The Solarfla
103.
by deadgrey19 12y ago
> Calling this the fastest switch in the world and then printing 5ns is misleading. The claim in the (original) story title is 110ns layer 2+ switching. I'm sitting in the summit right now, no details on what 2+ means, but, the targ
104.
by deadgrey19 12y ago
Timing in 10GE is sub-nano second. The PCS (physical encoding sublayer) runs at about 11Ghz (<0.1ns), so I think there's still a long way to go.
105.
by deadgrey19 12y ago
Indeed! It looks to me like there's two of them, so the potential is pretty incredible.
106.
by deadgrey19 12y ago
"Latency from 550ns to 2 microseconds" ( http://www.arista.com/en/products/7250x-series ) "Consistent latency as low as 350ns for all packet sizes" ( http://www.arista.com/en/
107.
by deadgrey19 12y ago
They said in the presentation (paraphrasing) "Evaluation units available late 2014, retail units available early 2015". My impression is that pricing will be similar to other high-end switches e.g. $20-30K.
108.
by deadgrey19 12y ago
Exablaze, today announced at the London STAC Summit that the company has introduced the world’s fastest network switch and application platform, the ExaLINK Fusion. The ExaLINK Fusion performs conventional layer 2 switching at approximately
109.
ExaLink Fusion – Ultra low latency switch(exablaze.com)
58 points by deadgrey19 12y ago|35 comments
110.
by deadgrey19 12y ago
If they're building C/C++ with it then this is a good option: https://github.com/Zomojo/Cake - which is really just a front end to make, but produces really fast makefiles. It was initially developed to allow
111.
by deadgrey19 12y ago
Thanks for the very considered response. I'm pretty interested in this because some of my recent work has been about designing I/O APIs/abstractions. My reaction is due to my experience which is that threaded programming is s
112.
by deadgrey19 12y ago
Hmm. I don't think I made myself very clear. I failed to mention that I'm advocating 1 process per core, not hundreds (thousands) of processes. Over subscribing processes to cores has the same effect of oversubscribing threads to
113.
by deadgrey19 12y ago
I don't write web applications, and I don't use much JavaScript, so it's very possible that I don't properly understand the motivation for this blog post. However, as others have said there is a difference between some i
114.
by deadgrey19 12y ago
I haven't used either in the last year or two, so a lot may have changed, although, in the past, many of the low end parts had a built-in (slow) PCIe Gen 1 block, so only wrappers were needed to use it, no licenses. IIRC, this was (one
115.
by deadgrey19 12y ago
I fear that this article misses out on one of the most important aspects of cheap FPGA boards, which is the software stack. Do not be fooled. There is no freeware, open-source, GNU, LLVM etc software in the FPGA world. You will need to use
116.
by deadgrey19 12y ago
Thank you! Perfect response. I'm a systems programmer, mostly working in VHDL, C and assembly. I use and write API's all the time. I've never used an "API" over HTTP before. I think the confusion comes from a slow d
117.
by deadgrey19 12y ago
But why not just use the script directly on that device? Since the interfaces to the "API" are things like Python, Ruby, QJuery etc, you still need your internet connected device to be running an actual scripting language. At whic
118.
by deadgrey19 12y ago
I really don't understand what's happening here. I think we use the term "API" and "server" differently. When I think of an API, I think of a library or something that I can link to. For example the POSIX API,
119.
by deadgrey19 12y ago
Fastpass is not "zero queue". They simply move queuing into other places. First, in the end host. When a host wants to transmit data, it must queue those packets and send a message to the arbiter to get a slot, it must then wait f
120.
by deadgrey19 12y ago
Incorrect. The scheduling only ensures that the queues in the network (from the NIC onwards) don't fill up. However, there queues are still there in the host and in the arbiter. The authors never measure or demonstrate that these queue

More ›