It is often interesting at the C level how things work. Ambitious people work their way up until they get to "the point" where they are either going in or going on. I got to watch this at Sun with Ed Zander as he made his way toward the CEO spot at Sun. When Scott "chose a different way" (which is what people seem to say when they mean "You are not getting the executive position you seek.") Ed left Sun and went off to become the CEO of Motorola. Marissa Meyer trying to become more senior at Google until it was clear she never would, and then on to Yahoo!.
It's a matter of math. The average public company CEO has 7-10 C level reports. When that CEO leaves, they can't all get promoted.
I mean it happens at all levels, it's just more news at the C level. If you want to be the VP of sales for example, and you're currently a regional director, there are probably a few other regional directors who want to replace your VP. Not everyone can do it.
I have been thinking of "Yishan-style CEOs" for a while now, and part of point is to deal with things like this better (including board of directors doing CEO searches publicly). Anil Dash once had a joke about it: https://twitter.com/anildash/status/893153626247626752
Cisco has been circling the drain for a long, long time (far longer than Arista). They are close in mold to IBM with declining revenue offset by managed profits from periodic layoffs and outsourcing. Cisco would continually find ways to cut employee perks.
They thought their thing was software consulting but they couldn't/wouldn't compete for talent with the Googles and Facebooks of the world. Note that Google really hit its stride in the early 2000s. That is a good 13 years ago.
People at Cisco knew even then. People would game the system by starting companies to be bought by Cisco and then quiting and repeating the process. A long time employee I knew set his price target at $30 and I think sold at least a large chunk when it temporarily hit that. My impression was that there were many slackers at the company.
A company mostly concerned with internal politics... much like IBM... presentations with org charts reminding me of my co-op time at IBM where people talked mostly about whom was the boss of whom... especially important since for Cisco, layoffs sometimes seemed to hit whole groups.
Cisco is circling the drain because their biggest customers finally realized they can build better products than Cisco using open source tools and commodity hardware. The big telecoms are all dramatically reducing spend and moving to proprietary in-house platforms.
Companies like Arista are selling white-box network gear at low margins because they know the market is commoditizing. Cisco got addicted to fat margins, but they have almost no patent advantage in most of their biggest markets. It gets hard to justify those margins after a while.
I'd love it if someone could expand on this a bit. My infrastructure knowledge sort of dies around 2005-2006, when I left Arbor Networks, at a job where we had to know the details of the various TCAMs in different line cards for Cisco Cat6k switches. From '05-'06 I did a fair bit of experimenting with network processors (oddball embedded CPUs with fast packet processing mechanisms). Then I got sucked up into application security.
The thing is, I've spent a career writing C code to parse, break up, and route packets for every conceivable protocol, in userland and several different kernels.
What I can't tell is: is my skillset here totally outmoded because of the new network infrastructure model, or are we in a new golden age for it?
(I am fine with either answer! I just want to know.)
A big shining place is in DDoS mitigation. We demo'ed some in-house built replacements for $VENDOR and found it to be cost-effective to do some more easily implemented parts in-house.
Based on some of my discussions with folks, I think people overestimate the cost-effectiveness of not doing some HW offload/out of kernel packet path stuff in specialized situations. Anyway, Amazon is looking for folks to work on their in-house DDoS mitigation pipeline per the reqs that float my way.
I'd say we're in a new golden age, after which we'll have a bunch of generic, open source implementations with very little need to engage in that type of development again.
This is actually happening all over the tech world -- companies like Google and Amazon are intentionally commoditizing key technology in their supply chain where there are a few powerful vendors extracting rents from everyone (see Cisco in the late 90s/early 00s).
Just remember that 50 years ago, Ethernet didn't exist as a concept. This stuff moves fast, so the fact that your skillset has been useful for 40+ years tells me that it's probably not going to be highly valued for too much longer :)
I can attest to OPs observation. Earlier in the day, I left detailed reasons for Ericsson's decline and laying off 25000 people.
Main observation is: Networking is getting commoditized.
I too have spent whole career writing C code and don't think there's a new network infra model. In past 25-30 years, it's matured enough to satisfy most of customer needs. I don't see growth here.
You're fine but more real networking is migrating to sw stuff on aws+etc clouds. (Note: not sdn, think old school Zeus, Nginx, haproxy and orthogonally c* and kafka )
Moore's law etc makes what used to be done in fpgas and asics, now feasible in clusters of VMs.
Yes there are exceptions but not that matter for masses.
I'll go one further and say that the white-box gear is better than the Cisco gear at this point. Had to deal with a network gateway at work last year which a) did not correctly implement STP, which we needed to hound Cisco to eventually patch, and b) hideously mistreated TCP connections so that they would be stuck open. I dealt with that second issue every day until my last at the company, I suspect that it's still b0rked if they haven't replaced it.
Accton and Detla and Quanta are whitebox. Arista uses off the shelf switch ASIC from originally Fulcrum then Dune, Broadcom and now Xpilant. The motherboards and all other hardware is in house. Not whitbox.
Stuck at 1GbE/10GbE and yea commodity will catch up. When $VENDOR is using off-the-shelf chip to do same speed for cheaper they expose themselves to whitebox vendors competition.
If there is no market or government pressure to level-up infrastructure (just optimizing existing network) not sure how $VENDOR will carve out any new revenue.
White box silicon is OK if you're deploying a campus, deploying a cloud where customers have no insight into the network, or situations where small packet performance doesn't matter. Everyone uses Broadcom (even though Mellanox spectrum is better), and Broadcom chipsets are trash. They're basically dump trucks that move lots of big packets, but can't do anything else.
There will always be markets that want vendor designed NPU's because small packet performance, versatility, etc. Juniper QFX 10k has shown that they can compete and beat Broadcom at their own game.
After watching Arista Broadcom switches drop traffic for years (they are lossy, sorry folks), I'm not touching them again. It's sad because they had something really special with their earlier products. I even owned some Arastra gear (yes, they renamed early on).
But isn't this the Valley Way? Where did Cisco's founders come from? Where did Juniper's founders come from?
This is why the Silicon Valley thrives: people leave ossified old companies and start new ones where they can bring their ideas to fruition more quickly.
>Cisco Systems was founded in December 1984 by Leonard Bosack, who was in charge of the Stanford University computer science department's computers, and his wife Sandy Lerner, who managed the Graduate School of Business' computers. (wikipedia)
(from what I understand, Sandy was the driving force behind Cisco throughout)
Until Sequoia's Don Valentine gets the VPs to rebel and have Sandy fired. Who ended up selling all his founder stocks for pennies, while the rest became billionaires...
It's strange to me that someone can be a good enough businessperson to get all the way to the executive level of a huge company, and still feel emotionally hurt that someone left to work at a competitor. That strikes me as a rather childish reaction to business as usual.
I was surprised at the reaction of my manager when I resigned from my first big kid job. I expected him to be supportive and happy for me as he had always been. Instead he seemed personally insulted as if I had betrayed him.
It was the right move for my career objectively but there seemed to be some resentment that I was "just in it for the money". Working a dead-end data analyst job at a company that is actively shipping its technical talent offshore was (and still is) not my idea of an attractive career path.
The HR rep pulled me into an impromptu meeting to find out what it would take to get me to stay. I told her I would need at least as much as the offer which was 40% more than my current salary. Her response was "do you think you are worth that?". I told her the other company thought so or they wouldn't have made the offer. Needless to say I didn't stay.
I wouldn't even call that mean so much as incompetent and desperate.
I'd maybe try to explain a little that my salary is actually set through a competitive process of interviewing so it's not a number I personally have to have believe in or not, and then maybe touch on why her comments smell of desperation and why that's not a strong negotiating position.
But I didn't say that "people evince childishness regardless of age".
I specifically said _jealousy_ is a trait attributed to humans as a flaw.
That has nothing to do with it being childishness or not. Otherwise it would be a trait somehow eliminated after childhood for a majority of adults, which it isn't since all of us (adults) experience it.
Furthermore, it elicits a sense of flight or flight, and causes a reaction, good or bad, I think. Again, it's a reaction; people in general aren't good at controlling those, nor do they last. A person can react as much as they want, but after a night's sleep, everything takes a different tone.
Our subconscious minds play a role that is probably undervalued. The jealous boss might wake up feeling relieved, invigorated.
It's referred to as a "childish" flaw because in children the prefrontal cortex is not yet developed enough to mediate base emotional responses like that.
Yeah, same feeling here as well. In his otherwise excellent book Shoe Dog, Phil Knight spent a paragraph talking about Rob Strasser, an early employee who left for Adidas, and about how it was a breach of loyalty.
Did you skip the middle of the book? Those two people went through a lot together, especially within the forum of facing-off with their competitors, one after another. Guy was integral to Nike and ultimately he chose to work for the "enemy." Think what you want, but either Phil Knight was too much of a douche for too long, which wore him thin, or he stabbed his boss in the back.
Seems plausible to me. I don't think it's an accident that the way we currently organize work looks like a giant primate dominance hierarchy. The people who rise to the top are going to be very good at that. (And hopefully good at business as well, but we can all think of counterexamples.)
TBH most men in the Valley are adult children, and always have been (looking at my mentor who's going on 70 now and was deeply involved in the early times of PC hardware). Childish behavior is hardly unexpected, and I truly do not mean any cynicism with this; it really is just a fact.
People may love to say "it's just business" when they do something that's hurtful to you; they'll still treat it as treachery when you turn the tables.
So, I totally read that title as "Cisco's feud with Former Star Trek Executive Turns Personal and Costly", thinking they misspelled Sisko from Deep Space 9. Been a long day...
>Mr. Chambers felt betrayed by Ms. Ullal, a former Cisco executive said. "To John, it was a relationship question -- 'Why would you do such a thing?' "
WTF. This guy is insane. If you're in switches, you're going after Cisco. There is no other way to build a company in that market.
I can't help but wonder if this is a well-disguised PR piece from Arista. While the feud itself seems likely, the whole article, down to the details, is heavily reliant on David vs. Goliath themes to paint Arista and Ms. Ullal in a positive light.
Given Arista's strategy of directly attacking Cisco's market share and "keeping quiet about it" you can bet they had similar sentiments about Cisco. Those details are missing.
Agreed, I hope that isn't the case here. Thanks for pointing out Cisco is part of her beat.
In the past several years I've become increasingly skeptical of news, particularly personal stories where the evidence seems slanted in a certain direction. There's usually another untold side to the story that remains untold as soon as the general narrative is established.
Sometimes they work out, sometimes they don't.