Jump to content

nikb

Ripple Employee
  • Content count

    733
  • Joined

  • Last visited

  • Days Won

    22

nikb last won the day on November 4

nikb had the most liked content!

About nikb

  • Rank
    Veteran Member

Contact Methods

  • Website URL
    https://ripple.com

Profile Information

  • Gender
    Male
  • Location
    Las Vegas, NV
  • Occupation
    Software Engineer
  1. We are always hiring. Do visit the jobs site.
  2. How to get a job at Ripple

    Apologies for not getting at least an email back. I’ll touch base with the team responsible and ask what we can do to make sure everyone applying gets a response, even if only to say thank you.
  3. This is viable given a large enough pool of validators and some assumptions about the possibility of any pair of validators colluding. It’s not safe to do yet.
  4. I'd be very interested in participating in that conversation, if for no other reason to listen and learn.
  5. I agree that validators are only really useful if they are seen as reputable in the eyes of the broader community. Sometimes (most times even) that will mean that validators operated by pseudonyms won't really pass muster. I'm ok with that.
  6. First: what Sukrim said. Beyond that, the presence of a large number of diverse (by any measure of diversity: by geography, by ownership, etc) validators will be a good thing, as the key consideration is choosing validators that don't collude. The more choices you have available, the lesser the chance that you'll pick colluding validators. Let's look at simple combinatorics. But first, note that I'm not suggesting that this is a mathematical model for choosing a UNL. I'm merely explaining why having many validators available is good. Assume we have 10 validators and want to pick 5: that's 10C5 or 252 combinations. Now, assume we have 25 validators; if we want to pick 15, we have 3268760 combinations. More generally, see this graph on Wolfram|Alpha which plots the function (X choose 0.8X). I somewhat arbitrarily picked 0.8 as a constant. Why is this important? Because it means that the more validators you have, the more UNL combinations you can make, which makes the network more resistant to an attacker that will target validators. "Wait a second," you may say. "Combinations are one thing, but what if you have 35 validators and know that 10 of them are colluding together but don't know which 10 and you want to pick a UNL of 20 at random? Huh?" Well, the statistical analysis would certainly be complex (and perhaps too long for a forum post but some of this is discussed in the whitepaper), but the basic intuition from the previous example would still be applicable and useful.
  7. The capacity question is a good one. Fortunately, we do have two things going for us: Capacity is increasing while the cost per unit of storage is decreasing. While this trend holds it will mean that, at a given point in time, it will continue to be possible to store more stuff for less in the future. And sharding, which will let the network, as a whole, hold complete history without any nodes having to do so. That's not to say that individual nodes won't choose to hold the full history. My own full-history server will continue holding full history even after sharding is implemented; I imagine that others will too, including Ripple (the company). Of course, this doesn't address the size of a "single" ledger. Reducing the size of the ledger data is something we're actively thinking about and investing time in doing and we will continue to do so. But right now, it's not a significant issue and I don't expect it to be in the near future.
  8. Codius is back!

    Soon™ hopefully! I need be able to catch some ZzZs on the Las Vegas to SF drive.
  9. The “UNL” is the (old)name given to the list of validators that the node’s operator lists as trusted. Basically, the admin of each server chooses their UNL but Ripple does supply a (changeable) default. As Sukrim says, additional, independent validators nodes do help the network, even if they aren’t trusted. With that said, while I do welcome new validators, I’ll urge everyone to stick with the default settings until phase 2 of the decentralization rollout is completed.
  10. Your remote keyfob may be somewhat secure against hacking, but this is not the concern here. The issue is digital signatures, and more specifically for our use case, elliptic curve digital signature schemes. The schemes we use are based on the notion of problems which are hard (where hard can be quantified mathematically and may even mean "impossible") to solve, but where solutions to the problem can be easily verified. What we use is, more or less, state of the art. But advances in computing, and more specifically, advances in quantum computing, could make it possible for quantum computers to efficiently solve problems which classical computers cannot. Some of the problems that can solved by quantum computers are the kind of problems that some of existing signature schemes use. This may sound like bad news, but not all hope is lost. The good news is that cryptographers have been hard at work on post-quantum cryptography for a while already and we do have good ideas on how to design "post-quantum" algorithms. As David said, we don't really have any great quantum-resistant digital signature algorithms that are suitable for our use case, but this is an area where there's a lot of active research, and we are keeping up to date with the latest. Also, even without an algorithm update, @JoelKatz and I have discussed ideas that can help us stay a step ahead of quantum computers. Maybe a topic for a future devblog post? Stay tuned!
  11. New Ripple Headquarters

    Have his people call my people...
  12. New Ripple Headquarters

    That's awesome. I love it. Also, dibs on a top-floor office with a view!
  13. XRP Escrow????

    As I said, mistakes happen. We are only human.
  14. XRP Escrow????

    Ahaha, that made me . What a great joke.
×