Our startup EntryDNS a free DNS management service | Hacker News Reader

Our startup EntryDNS a free DNS management service

entrydns.net

31 pointsclyfe25 comments

If it's totally free, how will you make money? Where's the business?

We'll be working on some premium features which will be charged a lot lower than competition. Most core features will be free tough! So we're going towards a freemium model.

Can you please elaborate on the features you guys consider premium?

> REST API for power users to update/change DNS records

Is there documentation for this? I was only able to find how to update an IP address. What about adding domains/sub-domains through the API, etc?

Is not yet documented, but yes, our REST api supports all operations on zones and records and is ActiveResource-compatible

http://api.rubyonrails.org/classes/ActiveResource/Base.html

Will e-mail redirection be among those premium features?

We have not decided on this one, but the service will definitely have this feature.

You should add support for SRV records. I know i personally couldn't use this unless you did.

Absolutely agreed. We are working on it at the moment. Give us a few days and we'll implement this.

SPF too.

Since an SPF record is actually a TXT record, they already support SPF

Looks good, although the previous mentioned point of your revenuestream is a big questionmark, I'd rather pay 1USD/domain or whatever to know it will stick around.

I tried looking at your ns-servers for the screenshot-example(100armstrong.biz) but that seemed fake. Your own domain are using:

answer: entrydns.net 3600 NS ns3.serveriai.lt answer: entrydns.net 3600 NS ns4.serveriai.lt answer: entrydns.net 3600 NS ns1.serveriai.lt answer: entrydns.net 3600 NS ns2.serveriai.lt

Checking what serveriai.lt is makes me wonder how stable that can be. Not beeing from .lt its hard to know if thats a thrustworthy provider. I would suggest you put a few examples ns-records for the domains that will be hosted by you, and perhaps an action plan for failures at any of those, or a nameserver upstream. A good addition might also be to show how geographically distributed your ns-servers are, and perhaps give users a few options so they can make sure theres always a nsserver within the same country or continent as most of the users.

Your upcoming feature-list looks really interesting and something that could be well suited as a paid-for-option.

Good luck!

Edit: got some more ideas after posting

I quickly setup an account to see how many name servers you are offering. There only appear to be two primary name servers that are within the same netblock. Both IPs appear to be announced by the same ASN from the same location.

So your clients are one medium level denial of service attack away from a complete outage.

Doing your own IP anycast requires some heavy investments but sometimes you can work with providers that have multiple locations to setup anycast in their IP block.

Denial of service attacks are something you just deal with when you offer low cost or free services like these. You are going to have a lot of unsavory types register and start hosting sites from your name servers; and even after you terminate the record your name server still has to deal with responding with NXDOMAIN.

Thanks for the feedback.

We're aware of that and we're working on it. The first goal is to create a well-tested and stable platform which we can improve over time with a help of feedback like yours. We do not expect people to start moving their critical services onto our NS servers.

Thanks again!

A DNS service with any type of free tier of service does not work. They have been tried dozens of times before (including by me), and never work out over the long term.

Your number one problem will be abuse. 50% of your user base will be the hacker/hobbyist types you expect, the other 50% will be people who can't get DNS service elsewhere (MLM/Ponzi schemes, child porn, jihadist forums and propaganda sites). These types of sites will generate tons of inbound complaints which you need to deal with, as well as massive DDoS attacks which you will need to sink to avoid service disruptions (think $100k+ in hardware plus bandwidth costs).

If you are crazy enough to move forward, you should reach out to dnsbl.org, its a shared database of domains to deny service to.

Nice too see a free alternative, even though I suspect it won't feel reliable in practice for serious sites, though I'm hardly an expert on this topic so I better leave that to experts.

What I could say though is that I'm after 6 months of usage very sold on DNSimple, those guys just rock and it's a few bucks per month - read: nothing - and updates records fast-as-*, instant support, even for obscure things as "the great firewall"-issues, etc. For those that want to pick a safe card.

Why would anyone sign up? No SSL and no security rationale talked about anywhere within the site. DNS isn't just for fun and games, the site doesn't exude much confidence in security as a development component.

What will your service have over dns.he.net?

Well some things I can think of are: friendly user interface, simple API and complete REST API compatible with many programming platforms, lower TTL values, instant updates, and several advanced features we plan to release soon.

A few thoughts:

* Your minimum TTL is 60 seconds, but you only learn this by trying to set it to something lower

* I actually like Hurricate Electric's interface better: there is much less clicking to do. Their updates are also instant.

* You lack IPv6 support (both AAAA record types and, from what I can tell, IPv6 connectivity). This is really important and would be a blocking issue for me.

* The REST API is something HE doesn't seem to have so good job there.

* You don't have HTTPS available, yet you have a authentication and management system. This is broken and would be another blocking issue for using your service. <rant>Browsers should display a big fat warning next to each <input type="password" /> field that is not served over HTTPS. In 2011, not having HTTPS set up should be a #1 bug.</rant>

Otherwise, looks pretty cool. If HE ever changes their services, stops being awesome, I will keep you guys in mind.

True. Totally agreed. We are aware of the mentioned issues and will be solving this in the near future.

Thanks for the feedback tho!

No problem and best of luck.

Please give any constructive criticism!

Here's one: You need to tell us why we should expect EntryDNS to exist in six months.

With no apparent revenue source, I have to assume EntryDNS will drop off the face of the earth at any moment, never to be seen again. That's not something I'm prepared to spend time putting a critical piece of infrastructure into.

I'm thinking exactly the same thing and wouldn't use this service for anything except testing. I'm assuming that your business model is freemium but even so you need to write a couple of reassuring words about the mium part of your business to reassure concerned potential customers.

The same answer I gave to eroded: we'll be going towards a freemium model - keeping core features free, and for advanced "enterprise" needs, go for fees lower than other services. We plan to stick around for a lot of time to come.

Reply on news.ycombinator.com