Hacker Newsnew | past | comments | ask | show | jobs | submit | naww's commentslogin

http://en.wikipedia.org/wiki/Jumbogram

> An optional feature of IPv6, the jumbo payload option, allows the exchange of packets with payloads of up to one byte less than 4 GiB


People are praising the transparency of this report, but I am not sure I agree because of this point. when I read the report, I had to stop to think when I read the part about packet size to conclude that they had to be talking about an IPv6 packet using the hop-by-hop extension for fragmented packets. That is a special case, because you don't actually know the length of the packet until you receive the last fragment.

As a consequence, fragmented ipv6 packets are error for use in DoS attacks. This is not a "weird" occurrence, but rather an expected one, and since end points are not required to accept such huge packets, I am surprised Cloud flare want already doing all it could to advertise to upstream sources that IPv6 fragments longer than a much smaller than 90K should be dropped, at least if rooted to their DNS. I am also surprised that when their software came up with that kind of a response without first validating that it wouldn't cause the exact memory problem it did. Rules on v6 fragmented packets that can't match on a single fragment are inherently dangerous. It is only reasonable to have safe guards already in place for them.

I am also not sure this is really a bug in Juniper software. I imagine the memory problem only shows up with high traffic and in the midst of a DoS attack. That is kind of a given when you put a rule like that in that kind of a situation.


Yes but they were still seeing packets bigger than the MTU of Ethernet (or Sonet or whatever other layer 1/2 tech they're connected to the rest of the net with). It doesn't matter what higher level protocols can handle.


They could've been fragmented IPv6 packets. Or it could've been a bug in their profiler.


which is precisely why it seems like lunacy to roll out such an asinine firewall rule to every router. if there was ever a time to "spot check" a change, this was it.

they didn't. and they paid the price. good on 'em for the quick and honest post-mortem. regardless, it was a dumb move.


You are joking right? The packet size at the higher layer is what they were matching against. The size of the layer 2 packets is irrelevant.


Maybe, but nothing in the the rule they showed hinted it was not at layer 3 (For IPv4 )


It is at layer 3. IPv6 is layer 3.


If it was IPv6, I'd assume the routing rule on their blog contained IPv6 addreses, not IPv4 addresses, even if the blog faked the IP addresses.


Perhaps then you aren't aware that IPv6 stacks can reach IPv4 addresses, nor that IPv6 packets are a popular way to compromise systems that support both IPv6 and IPv4, because the IPv6 stacks are not as well hardened.


I was thinking about getting "sheeple media", dot com and trademarks but it is a bit condescending if you know what I mean.


What you provide seems very very trivial for any web programmer. How you explain that this service adds some value for a company?


Tweaky is for small business owners and non-technical people.


Anytime someone starts carrying a gun around and acting like a cop while being trained and uniformed as a cop, there's going to be problems, if you film them.


I don't see why they use civilians to look and screen videos. Hope they have a watchmen for watchmen and informed officers who know about taping police on work.

Isn't this bit off Hacker News 'thou?


So you would need to write your own Foreign Data Wrapper and implement some mechanic to tell it if the SQL code is meant for first or the second interpreter?

Like: SELECT * FROM redis_db0 USING_REDIS limit 5; And Postgres would just pass that"USING_REDIS limit 5" through to Redis? Then Postgres would pass that table or array? back to caller.

Wait, wouldn't this also need modification for DB engine itself?


Today was my first real usage of PostGres (moving one of our assetdb from sqlite to it, and although the update of the db is 5 time slower - well I haven't really done well there, the query later (which is more important) is from 1.5sec to 300msec on one recent example - about 100,000 assets for a video game).

I've also looked briefly about fdw, as a coworker of mine is trying out mongodb for our level editor... Looked at all implementation and read the every fdw code there (except oracle - as it seems most complex, but for good I guess - it checks the pgsql's AST (is that how it's called?) more deeply).

As for redis, maybe this line is relevant:

https://github.com/dpage/redis_fdw/blob/master/redis_fdw.c#L...

yes, for certain cases (I guess "select ") it does "keys ". I'm also trying out redis for our distributed caching in the studio.

So PostGres is my new favourite thing :) - the fdw looks easy to start, but probably hard to master it.


> Today was my first real usage of PostGres

Then let us welcome you with some pedantry! Postgres or PostgreSQL, but never PostGres or--FSM forbid--PostGreSQL.

More seriously,

> fdw looks easy to start, but probably hard to master it.

Yep, like a number of Postgres features, such as user-defined types and user-defined aggregates. It's a great system, and not nearly as stuffy or hard to use as you may have heard.


Ah, I will be careful how I write it :) Thanks!


Still no way to contact human for a problem. No feedback or any kind of acknowledgement about anything.

And my only problem with Google only appears when I'm not logged in. The irony.


Depending on the product and the nature of the feedback, you can often go directly to a team lead on Google+ and get some kind of reply.


Meh. Not getting any kind of account just to give feedback.



> Supercell attacks the $ 7-10 billion mobile games market,

So does every other mobile game company. Why this one is better than others doesn't open to me by reading this blog post.


Founders are game industry veterans.

A recent proof that Supercell founders know how to execute is that they changed their focus from Facebook games to iPad and have released two successful and polished iPad games in a short period of time. Both Clash of Clans and Hayday are hitting top grossing lists. And they have a third polished title, Battle Buddies, coming out soon.


The author is investor in Supercell ;)


But what about "release early, release often"?


Is there a difference between that and "release early, release often, release broken"?


Nope. Don't be afraid to release broken. As full quote goes: "Release early. Release often. And listen to your customers".

Trying to perfect a product by your self is hard and other people always find a genuine way to broke it. I believe that it is more important to be first out, first with samples and first with some kind of a product to sell that improves with time.

Thoughts?


It depends. What sort of broken do you have? What is your business? What is your scale if this is additional feature on existing product.

Usability? How broken?

Performance? Might be ok if you have no initial customers.

Dataloss risk? Depends on the business.

Security problems? What is the worst case - personal photos (whose), medical data. Legal risks?


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: