https://www.washingtonpost.com/news/worldviews/wp/2018/05/29...
1,168 karma · joined February 11, 2015
https://www.washingtonpost.com/news/worldviews/wp/2018/05/29...
Good example: The SRX300, a desktop router/firewall implemented on a MIPS based Cavium Octeon (both control and data plane) can handle up to 256K IPv6 routes: https://www.juniper.net/assets/us/en/local/pdf/datasheets/10...
The Trident series SOC's can do 8K by default, and 64K with ALPM enabled: https://www.juniper.net/us/en/local/pdf/datasheets/1000480-e...
>(I mean, you only use /127's for point-to-point links that don't need link-rate switching anyway)?
/126's and /127's are often used for PTP's for peering/transit. Line rate is table stakes. Bear in mind that the gear used in these situations has plentiful lookup memory (SRAM, RLDRAM, HMC, or HBM).
Why can't you subnet the /64? CenturyLink isn't doing dot1q tagging to you, so you must have a router with tagged interfaces on your network.
Edit: Imagine this is because you want to use SLAAC.
From the article: "Provides a very simple mechanism for a single host or interface to be able to run 2^64 virtual machines"
That equates to AT LEAST 2^64 NDP entries. And when the L3 domain is scoped to that of a single hypervisor, router redirects are not relevant.
NDP is the abominable child of ARP and an IGP and doesn't scale well. Using huge address spaces everywhere is a good way to let it crush your network.
And it's "losses" :)
A hard ratio seems fair minded, but seemed to have failed here because I think there was a misunderstanding around compensation. Compensation is a product of many variables including value created for the company and supply/demand for skillset/talent. As a business climbs the value chain or seeks to hire and compensate those that create greater amounts of value, and it maintains a compensation ratio, it must seek to increase the value produced by the lowest earners (or lose business to competition).
Looking back, often no talent has been better than poor talent (or those that actually decrease group productivity). Huge opportunity cost with bad hires, and even more challenging to move them out to somewhere that they would be happier/more effective/less disruptive.
[ Edit: Nevermind, found the data sheet: https://www.netronome.com/media/documents/PB_NFP-4000.pdf ]
One remaining question - how big is the space allocated to lookups/LPM?
It's the abuse of immigrant labor that's wrong.
Take a long look at the companies providing feedback around your time off being troublesome. Likely, you wouldn't want to work with them.
Control-plane/management-plane software is often written by these companies or by contractors.
NICs are another story. MSFT and AMZN both build their own NICs using either FPGAs or custom silicon.
Google does operate (lots of) vendor gear, particularly at their network edge.
However, did you note whether or not these efforts may have been positive contributions to the healthcare of humans? I would imagine gathering health data from a large population has the potential to yield remarkable results.
I don't think it's any defense - it's plain fact. You're going to have to interopate with other networks. BGP is the protocol franca for doing so (until another rolls along). You can draft a spec, or add capabilities to BGP, but at some point your network needs to speak the same language as others or it's just an island.
>There are BGP alternatives on deck. What we don’t need is high profile network engineers speaking out against them and in favor of the status quo. The truth is, networking is the one part of the industry that isn’t improving at the same pace as the rest. And all the improvements that are happening seem to be related to single links and switching, whereas the defense of networking dogma is related to routing. According to the holy bible of networking, switching is easy, routing is hard. Routers must be memory restricted with small memory footprints to artificially create SKUs. Routers must operate on multi-minute timers rather than seconds. Building routing tables is an insanely intense and time consuming process. Blah blah blah.
I don't know where you're seeing this. The rage around Clos networks has sent everyone after small table SoC's. Any one operating at scale is running some form of commodity (aka, BRCM) chip as a router.
And when you get to routers, the Jericho (purchased by Broadcom, found in the Cisco NCS) or the Juniper Paradise (found in the PTX) are both capable of taking in internet routing tables using either HMC or off-chip memory in a pizza box form factor and in a fixed pipeline chip. Your statements aren't aligning with the reality of the platforms available today.
>Another smell that should make this obvious to everyone is that almost every major company has had high level engineers make some statement of how network companies are ripping you off, and explaining how they bought white box network solutions that route layer 3 at wire speed, converge rapidly, and have tons of memory.
This last part reads like the grass is always greener, and doesn't consider a lot of the hard work and time that goes into developing network hardware.
First, conflation: White box can entail NPUs or ASICs, which have different implications on FIB memory - the majority of these major company (read: FB, MSFT, GOOG, AMZN) use ASICs with some on-chip memory (in the dozens of MB at best). Convergence speed is primarily driven by the network operating system and it's protocol implementation (aka, something you need to write, or contract out to a vendor)
What those high level engineers don't tell you is how hard it is to create white box solutions. And they aren't vendor free paradises either - you have to sign a NDA with the chip manufacturer and use their SDK. All of those things like implementing counters, make before break, prioritization, control-plane and data-plane interfaces are up to you. And when that's all done, you need to perform countless hours of testing.
Why not take the standards approach and draft a protocol spec? At some point, you'll need to interoperate with those horribly dogmatic vendor hugging networks that serve the majority of traffic on the internet.
>who typically have dogmatic vendor based training rather than a “network science and theory” education that leads to innovation.
I've seen this as well, and it can be frustrating. Part of me believes this is a defense mechanism used to prevent developers from pushing problems into the network. The other part believes that network engineering has drawn individuals to the field based on earnings potential rather than interest in the work, and the talent just isn't there. Both are anecdotal, so grain of salt and all.
If owners are pocketing more revenue, in lieu of distributing it as profit sharing or employee salary increases, that would seem to indicate the salary bargaining position of current and new employees is not strong.
Regarding meeting them in the middle, I would put together a presentation that describes how your algorithm works at a high level. I'd do your best to split the balance between being transparent and focused on the customer, and not divulging what you consider to be differentiating parts of your implementation.
If you get push back for the above, I think you're dealing with either brain rape, or some pretty unprofessional contacts in your customer's organization. If it's the latter, you should do your best to navigate around those contacts.