For my non-techy friends/family: in case you don’t understand exactly what I will be doing next year. (Written while on the plane to New York in FAQ format, because I had nothing else to do.)
Q: Hey Ryan! So what are you going to be doing next year?
A: I’m starting work at a startup, working in the wild world of 10 Gigabit Ethernet Switches.
Q: 10 Giga-whadjyaba-whaaa?
A: Yeah, something like that.
Q: Wait, so what is that? You’ve literally given me no information about what you are going to be doing.
A: Short answer: I am gonna help make the internet fast!
Q: And the long answer?
A: Well, since you asked…
The internet is a giant network that moves a lot of data around. You can kind of think of it like the US Postal Service. Whenever you go to a website, you are sending a letter to them and getting a response. So when you want to go to Facebook.com, you are essentially writing them a letter saying “Yo Facebook, wassup?!”
Facebook then sends you a letter in response: “Nothing much, man. What do you want?”
Your computer: “I want my friends’ info, and the relationship status of that cute girl that I just met.”
Facebook: “So typical… wait a moment while I get that for you”
Facebook: DATA
Facebook: DATA
Facebook: 1000101101101
Your computer: “Ok, thanks. Bye!”
Facebook: “Adios.”
Now imagine if that conversation was hundreds of letters longer, and you start to get an idea of what a single webpage request is like (except we call them “packets” instead of letters).
Q: Uh huh….
A: So here is where the USPS analogy comes in. In order for your letters to get to Facebook (or whatever website you are visiting), they have to do a lot of traveling. One of the things that helps your letters get to their destination are routers, which you can think of as the large letter processing plants that look at millions of letters, figure out where they are going, and send them to the next stop. These industrial machines are giant behemoths.
Anyhow, once your letter gets close to its destination, it gets handed over to a switch. I like to think of a switch as a mailman. It has your letter, and just needs to figure out which house (i.e. server) to deliver it to. In large networks (such as Facebook or Google - think of these like large neighborhoods), you have a lot of servers, and thus you have a bunch of switches working together to process your request.
Q: So what you are telling me is that you will be making mailmen?
A: Yep, pretty much. Giant mailmen that can send letters reallllly quickly. We are talking about 600 nanoseconds fast. These switches are sending your letters on even before it finishes reading the address; the packet is exiting the switch while it is still entering it. Things also get a lot more complicated than it may seem. A whole host of things can go wrong: for example, your letters can get corrupted while in transit, and the switch has to figure out what went wrong. System administrators (the people that buy the switches and make sure their part of the network is running smoothly) also like all sorts of tools so they can figure out what is happening in different parts of their network.
Q: So whats so cool about Arista Mailmen—I mean switches?
A: There is probably a lot more to this question than I understand at this point, but here is why I think Arista is already successful/will be even more successful (in no particular order).
1. They are fast. The financial market loves the fast Arista switches because it means that they can place their trades even more quickly.
2. They have a bunch of smart people working for them, here are a couple:
a. David Cheriton: Co-founder, Stanford Professor, Baller. He has 8 great startups behind him, making him a nice billion and a half dollars.
b. Andy Bechtolsheim: Co-founder, Founder of Sun Microsystems, Technical Mastermind. Everyone who meets this guy can’t stop talking about how smart he is.
c. Jayshree Ullal. CEO, super-savy business person (who also knows more about networking than I will for a long time). She used to lead the 10 billion dollar division of Cisco’s networking business. Also named in a bunch of top lists for being amazing/smart/awesome.
d. A whole bunch of other really smart people, many of them from Cisco. These guys are cream of the crop smart, so all I know is that I will have to hit the ground running once I start work next year.
3. A better quality infrastructure: the Arista switches are built over their own operating system, which was designed to be modular, easy to use and update. This means that it is easier to innovate and develop features when coding on it.
4. It is cheaper. Yep, the Arista switches (from what I can gather online) are cheaper than the competitions at roughly $26,000 per switch. Feel free to get me one for my birthday :)
5. Other innovations.
a. The designs of these switches is just clever, and make the life of sysadmins easier. For example, the direction in which the fans are facing is important in a data center (there are hot aisles of exhaust, and cold aisles of air intake). Arista has got you covered there.
b. Other niche applications. For example, some networks need switches with deep buffers (i.e. mailmen that can carry a lot of letters at the same time). Arista makes such a switch, and it was used in Vancouver this last winter to allow live streaming of Olympic coverage. Pretty neat!
6. A bunch of other stuff that I don’t know yet. But I’m sure I will learn it in this next year!
Hopefully this was easy enough to understand and gave you a bit of info about what I will be doing next year!
Awesome online drawing application
Interesting fact: it was made using HTML5—not flash, like all of those other online apps.
Above image was created by my sister, as I have no drawing abilities.
I’ve been following the “No Flash on the iPad” stories fairly closely over the past couple weeks. Most interesting to me is the pundit’s takes on why Apple has chosen to neglect a de facto standard of the web. There is Apple’s claim that flash is just not stable/cpu efficient, Gruber’s perspective that the computing giant is trying to change the baseline standard of the web, and a hundred other theories flying around cyberspace. I don’t disagree with any of these explanations. However, I do feel that none of them are compelling enough to be the real reason why Apple is blocking flash from the platform.
My logic:
Other Notes:
