Hacker Newsnew | past | comments | ask | show | jobs | submitlogin
Oracle buys Dyn (oracle.com)
229 points by tribaal on Nov 21, 2016 | hide | past | favorite | 126 comments


Does Oracle know that people expect that everything Oracle touches will turn into garbage?

I mean Oracles image is absolutely terrible among everyone but a small subsection of large companies. Arguably it doesn't seem to matter, Oracle isn't going out of business any time soon, but I can't imaging that their brand isn't suffering to some extend.


Somewhat recently underwent an acquisition by Oracle. A very few meager retention bonuses were given out (on the order of 5% of engineers). An immediate exodus of most of the engineers, including those offered retention bonuses, began.

It seems counterproductive, but the trick was this: Oracle gained ownership of our contracts with other companies. Now, you might say, how the hell are you going to keep our products running if everyone who knows anything about them has left and they're all falling apart?

The answer is: this is upside for Oracle, not downside. They're big enough that they can ream customers by forcing them to pay for overpriced consultants to fix the problems Oracle itself caused.


But aren't the customers going to leave Oracle in the long run?

I'm not saying it's not working, because Oracle is clearly doing fine. It just seems rather short term, and really unethical.


Switching costs to individual decision makers are perceived as high, justifiably.

Imagine you're a decision maker, and you have the chance to either (a) stay on with Oracle or (b) go with some set of startups that provide everything Oracle does. Which is, to be clear, a whole fucking lot, even if it does them badly.

Best case scenario (which is hard to pull off), there are no hiccups and you end up saving your company millions or tens millions of dollars per year. You get some reward for it, but definitely not one ongoing or even close to commensurate with the value you just provided.

Most likely, you run into significant hiccups. Maybe one of the startups you chose was a bit optimistic for what necessary features they could deliver in time for the Oracle switchoff. Maybe it's unreliable and all the users hate it. Maybe it takes way more time than expected to retrain the staff to use all the new things. Maybe Oracle drags its feet about helping you stop being their customer. Individually these are all perhaps evitable, but you're likely to run into some issues. Then you have egg on your head. And even if Oracle was just as bad in all those things, because you were the one who rocked the boat, the blame is on you, instead of "well, that's just the state of life."

You might say, "well, why not do things piecemeal?" And you can try that. Indeed, you are ethically obligated to employees and shareholders to do that instead of introducing huge risk to everyone because of an understandable anti-Oracle vendetta. But the way Oracle structures its sales and legal strategies makes it much less financially beneficial to you to do that than to cut them off whole hog. So you scale down the risk, but the potential benefit scales down even more so. And of course, who knows if Oracle won't just acquire the company you switched to a year or so down the road. And then you're stuck where you were originally.

I think, practically, once Oracle has its claws in your company, you're stuck with the parasite. The only solution is to categorically cut off Oracle from consideration for any vendor-supplied solutions. If it acquires one of your vendors, you make it a priority to transition off of it as soon as possible. If there are no other options, you start a new company to fill that niche.


No, they understand deeply how much it costs to transition to an alternate solution and price accordingly.

When I sold computers at retail in high school in the 90s, I'd hang out near the return desk and deal with angry customers. Angry customers actually want to attract somebody's attention (biggest problem a salesperson has) and needs a problem solved.

In my case, the solution usually involved a better computer and a service contract. Literally 9/10 times, once you spoke to these folks respectfully and calmed them down, they were the best customers. In Oracle's case, they upsell all sorts of crap and end up in a position where they can set the agenda because with the money involved, things must work. The IT decision maker becomes dependent on them.


>But aren't the customers going to leave Oracle in the long run?

as Russian saying goes: "you can't run away from inside a submarine."


I like this saying, but can't track down a reference to it. Pointers?


The life in Russia is the great source and reference for it :) I really don't know that is the original source of it. That is what google brings, first link exactly discusses your question (all in Russian) :

https://www.google.ru/?gws_rd=ssl#newwindow=1&q=%D0%BA%D1%83...

some say it is from a song, some - from a joke.


>> But aren't the customers going to leave Oracle in the long run?

Maybe. Some of them won't/can't switch.

The goal is you buy a services company for $X with Y customers spending $Z per year. Doesn't take long for Y * $Z to be much greater than $X and you just string the customers along until they leave.


The trick is to acquire services where the data is valuable AND hard to export (think Student Information Systems or other enterprise software) -- so in the short term it's cheaper for the client to pay for consultant than to pay for migrating the data to an alternative.


> But aren't the customers going to leave Oracle in the long run?

In my anecdotal experience (n=1), yes. At my workplace, every Oracle product we use besides Java (that I know of) is being replaced for new projects and re-engineering efforts. Oracle BPM is getting dropped for Activiti, Oracle Database is forbidden for any project that will be deployed to AWS (due to licensing costs) etc.


I guess people already know about what to expect from the story of Sun.


It's because that "small subsection of large companies" includes the people who have the money and buy the stuff. It's not very useful to dismiss them. Oracle knows that their image among techies is absolute crap. They don't care. Why should they? There's an actual business lesson to be learned from them, for those who care to try.


Not that I'm an Oracle fan, but they do continue to develop MySQL. Both in terms of features and performance.

Oracle still invests heavily in the development of the JVM (graal and truffle aim to advance the state of art in JIT compilers)

So it's definitely not all bad (virtualbox is also used by many but unlike MySQL I don't follow the development)


MySQL is rubbish compared to many things (seriously. It's kind of amazing), although I won't discount how useful is, and how much Oracle has made the situation better.

The JVM stuff and virtualbox are both pretty good. I don't know how Oracle has refrained from mucking them up.

But if there's any product that you actually have to pay for, Oracle has screwed it up.


The single biggest thing we're getting in Java 9 is the ability to better run different jars on the same JVM.

Meanwhile, everyone and their dog moves towards containers and similar isolation technologies.


Not really a fan of oracle at all, but I have to congratulate the Dyn folks for how far they came from a dormroom server at WPI. I had the pleasure of spending a short summer internship there, and it was a formative experience. They had not only great engineering but also some of the most impressive sysadmin / ops people I've known. Especially the network stuff is a dying art these days.


What is WPI?


A tiny engineering school in Massachusetts. I'll spare you the marketing-heavy website, but It's an awesome place with a small-town feel, where nerdy pursuits are encouraged and helped to flourish into professions and businesses.


Oh sure, the acronym through me. Thanks.



"We Produce Idiots"

Source: went there in the early aughts, loved it (the teachers, the environment, the people... not so much the administration)


Hah, that one is new to me! :-)


Guessing "Worcester Polytechnic Institute".


I don't see this as a good thing at all. Dyn themselves are a good enough company, but Oracle's culture could easily leach in.


Indeed. My first reaction was "Oh <insert swear word of choice>, there go my domains".


1. Symantec buys LifeLock 2. Oracle buys DYN 3. Oracle buys massive protection policy for DYN via LifeLock 4. DYN gets hacked 5. Oracle submits massive insurance claim 6. Symantec stock tanks 7. Oracle acquires Symantec dirt cheap


For everyone currently rushing to move their DNS records away from Dyn to alternative providers; don't panic.

Moving your DNS isn't hard, but is basically a hit-or-miss: it either works or it doesn't. I'm working on a tool to help monitor DNS changes that can help in migration scenario's by letting you know if any unwanted DNS records have changed.

I'm looking for any feedback you may have, you can sign up for the beta here: https://dnsspy.io/

(Ps; it seems shameless to plug a new service in a thread where a company just got bought, but I truly believe it can help ease the migration pain some may be feeling. Feel free to downvote if it makes no sense here or if it's just too blunt. Apologies, in that case.)


DYN did a lot more than just Authoritative DNS. They did monitoring that then would facilitate a failover to different hosting environments, creating a quick active active or active passive failover depending on your configuration. They offered this with a complete web GUI that had permission controls that were granular so newer techs or less technical employees could login and manage just a single record or series of records.

Not to mention their solutions are very fast.

I'm not aware of any other dns host that does all of this.


Absolutely!

This only applies if you used Dyn as a DNS provider with DDoS protection, not if you've used any of their failover capacity.

Having said that, I also don't think people should rush to get away from Dyn. Take your time, evaluate the buy and see where the service goes in a few months. If nothing changes for you, why even bother changing?


Route53, Azure, and NSOne have it already. CloudFlare is about to have it too. Google Cloud solves this with global load balancers if you're running in their cloud.

There are several other DNS providers that have varying levels of this functionality. There is nothing unique about Dyn except their pricing.


>Moving your DNS isn't hard

I bet Oracle wouldn't bother buying Dyn if it wasn't hard, especially for big customers.


DNS is easy, mostly.

Most users just don't care about DNS, so they use their domain-registrars' offerings (which are frequently terrible). The people who do care seem to pick their provider based on either price, or on features such as non-standard record types (the various different hacks to allow top-level CNAMES, or the ability to return different results on a GeoIP basis.)

Once you start relying on these special things moving becomes hard - because different providers either don't offer them or don't offer them the same way.


looks at a thousand domain names across dozens of foreign registrars, running on tinydns, with views that may or may not work anymore

Welp, at least I'm not on Dyn!


I think we've interacted in the past my service is the reverse of yours - storing DNS records in git repositories:

https://dns-api.com/

I picked up a few new users at the last Dyn DDoS attack; I'm actually surprised many (Dyn) customers moved away entirely. To get redundancy users should add their DNS records to, say, Route53 AND Dyn, not just jump-ship to a new provider.


Note that signing up puts you on a waitlist and doesn't, immediately at least, get you onto the beta.


True, I should've clarified that.

Those who've joined the beta waiting list (by entering their e-mail address on the homepage) should get in within days. But I also don't want to rush things and break functionality, so I'm still being careful here.


I actually appreciate the structured scaling and it's probably a good choice. Seems like a useful service, for sure!


Well, this will no doubt go well.

I'm so excited to see what ...improvements... oracle will make to the Dyn product line. I mean, just look at all the exciting ...improvements... they made at Sun, and all of the other (too many to name) companies they bought out!


As an employee of Oracle that sells Sun stuff, it really isn't all that bad :)

But we are all lol'ing at the parallels here. A number of our colleagues went to work for Dyn recently, too haha...


> isn't all that bad

So, how would you characterize it? Somewhat bad? Doesn't it bother you to work somewhere so universally reviled?


I would say that we all have become numb to the negative aspects of this job, which as you can imagine, usually include:

1) (rightfully) angry customers 2) painfully arbitrary and illogical corporate procedures 3) any sort of purpose or mission that our products fulfill

Speaking for myself, I think it's very important to realize that this job is one of the best in town, is still decently well paid, and allows you the down time required to work on "real" projects.

The ones who were capable of leaving have left. The rest of us are waiting.


How do they feel about this? Are they laughing like you are?


They are laughing yes, I don't want to speak for everyone at Dyn, but the people I know are laughing at how hard it is to escape Big Red!


How long until getting set-up on Dyn requires a downloader to download an installer that might put the Ask Toolbar on your machine unless you uncheck a sneaky default?

Fucking Oracle.


Bye, Dyn. It was nice while it lasted, but I think it's time for both of us to move on.


Oracle is currently reviewing the existing Dyn product roadmap and will be providing guidance to customers in accordance with Oracle's standard product communication policies. Any resulting features and timing of release of such features as determined by Oracle's review of Dyn' product roadmap are at the sole discretion of Oracle.

Sounds promising!


I assume that was sarcasm, right?


“Jump! Jump now!” — Kosh Naranek


Indeed.

I probably would have picked Cantrill for the quote myself, but I have absolutely no subtlety.


A bright future ahead for all customers!

*snark intended


I bet the first change'll be to replace the pricing page with a contact form/quote so that you can talk with a sales manager about your "enterprise" requirements. Not very exciting for start-ups.


Yep. It's funny, because it makes sense short-term. Why bother with the headaches of little startups when you can have a single "enterprise" customer paying as much as dozens or hundreds of them? It's easier to provide service to a a few hundred big customers than hundreds of thousands of little ones.

Of course, then someone sees the gap, and provides service to all those little startups, and grows from there. Until they get purchased by a big "enterprise" company and the cycle repeats.

I bet anyone competing with Dyn right now is pretty happy with this news.


I'm now looking for recommendations for another DNS provider. USA or Europe preferred. Anyone have any?


Some ideas:

Google's Cloud DNS: https://cloud.google.com/dns/pricing

Route 53: https://aws.amazon.com/route53/

Dns Made Easy: http://www.dnsmadeeasy.com/

Dnsperf.com is also a good resource to compare speed and reliability of DNS providers.


I've been using Route53 for the last 3 years, and I'll vouch for it. It's pretty sweet -- the UI is clear, but allows for a decent amount of configuration, and it's reliable. It also integrates with major registrars (GANDI's a big one) and lets you manage domain ownership via an AWS account -- and personally, I prefer unified bills.


I'll anti-vouch for Route53 for personal use. I use AWS pretty broadly, but with something like DNS, I like having a little bit more support if things go wrong. Amazon won't do much without a $upport contract.

It's fine for business use with a support contract.


Route53 is awesome. I personally wrap it in git - https://dns-api.com/ - and remove the per-query pricing to provide stability.


Not sure if it's the same service, but if you're using Google as your registrar, DNS seems to be included... would be surprised if it wasn't the same infrastructure.

I've heard good things about AWS Route53 as well.


Google support confirmed that their Domains DNS is the same infra as the Cloud DNS


Yeah... google domains' management tools are what got me to switch pretty much everything over to them... some domains you can move to them after registration, but can't get new ones for the given tld.

I can't say how happy I've been with their mostly out of the way tooling... the two times I needed support using the callback option worked out for me. Far easier than getting support for anything else google, but it's been great so far...

Even tried the dyndns for an a-record, which has worked well enough.


CloudFlare. It's free and bulletproof, even if you don't use their protective reverse proxy you can use their DNS service for free and it has a dynamic updating API.

Unlike other providers (looking at you Amazon with your handling of Wikileaks) they've also proven themselves willing to stand up for their customers, even free ones legally.


+1 for Cloudflare. For failover and such, I'm watching for how their traffic control and traffic manager features evolve. https://blog.cloudflare.com/cloudflare-traffic/


They will, however, acquire an SSL cert in your name for themselves when you use their DNS service.


What problems can that cause though?

(Alternatively - does anything stop me acquiring an SSL cert in your name, despite me not being Cloudflare and you not having an account with me?)


Cloudflare is not bulletproof. Had several times where cloudflare proxy plus my Nginx server got stuck in infinite loops making everything time out. No logical reason why. Switched to cloudCDN and 0 problems since.


Sounds like your app was doing something weird as I've not had any issues using a similar setup.

What's this cloudCDN thing? Can't find it on Google.


https://cloud.google.com/cdn/docs/

Check a checkbox, and your static stuff gets CDNed.

Seems a lot better than letting another service control you, perhaps put up CAPATCHAs, perhaps mess up your networking, etc.


> Seems a lot better than letting another service control you, perhaps put up CAPATCHAs, perhaps mess up your networking, etc.

Disabling the CAPTCHAs is trivial, just set the security features to low or off. This does basically the same thing in terms of "mess up your networking" though. All your traffic is just routed through Google instead of CloudFlare. No real difference there. You can also kick the CF security up if you get attacked.

Google's will however cost you a crapload if someone decides to run even a fairly primitive layer 7 attack against you. So you better just hope you don't get hit, unless I'm missing something excluding attack bandwidth?


Google does do some automatic DDoS protection for you: https://cloud.google.com/files/GCPDDoSprotection-04122016.pd...

I'm not sure if that addresses your concern, but hope this is helpful. Let me know either way.

I'm a PM on Google Cloud, but not working on either CDN or DDoS protection.


FWIW I like dnsimple . I've got domains on both services and I'll be moving my last stuff off dyn and onto dnsimple after this, no desire to be a customer of Oracle's.


NYC-based NS1: https://ns1.com


Ironically they're now down - https://news.ycombinator.com/item?id=13019858


I too am looking for alternatives. I moved to Dyn from ZoneEdit, I may look to go back.

Has anyone favourable experience with Route53?


> Has anyone favourable experience with Route53?

Yes, using it commercially(for work) and for personal domains. The pricing is incredibly low, and I haven't had an outage/hiccup since it launched.


Thanks, I might give this a go with a 'spare' domain to investigate it more.


Ugh, I definitely wouldn't recommend going back to ZoneEdit (now "easyDNS").


I didn't know they were bought out as well. Many years ago I ran my own DNS, perhaps I should look at doing it again.


I use a combination of DNS Made Easy and Route 53. Both have worked well for me.


This is a good thing to do. Rather than host all your DNS records at any one provider, split them across two.

That allows you to keep on running even if your provider suffers a DDos, or outage. (As Dyn famously did recently.)


We balance our domains across Dyn and UltraDNS. No complaints about UltraDNS so far. Has just worked, and they have a decent REST API which is nice (as does Dyn).


We were on UltraDNS for a few years. They started receiving nearly quarterly DDoS attacks and then tried to try to sell us DDoS scrubbing services. From there we moved to Dyn and have been happy. The DDoS that took out Dyn for a while would have take any single provider offline.


Yeah, definitely. The bigger takeaway is probably that you really want to have more than one DNS provider. Assuming whatever you're running is important enough to justify the cost.

If DNS ain't happy, ain't nobody happy ;)


I agree entirely.


namecheap work well for me.


We're in the process of migrating away from Namecheap. They've been having DNS issues lately, as recently as last week (16 November) their entire DNS infrastructure was unexpectedly down for several hours. [0] Took all our sites offline.

> $ host mydomain.com dns1.name-services.com

> ;; connection timed out; no servers could be reached

They're cheap, but have been experiencing some real reliability issues as of late. Other registrars offer pretty similar pricing and don't suffer from as frequent outages.

We were hardly the only people affected. Check their Twitter[1] for other people complaining of DNS outages.

Also, their new admin portal redesign sucks. Tons of white space and managing domains now requires many more clicks than it did before. Their zone management interface also frequently fails to update the zone.

Gandi's zone management interface is a million times better.

[0] http://status.namecheap.com/archives/28960

[1] https://twitter.com/namecheap


Hate to pile on, but the admin portal redesign was incredibly bad. So many things that were logically organized together in the old design are now spread across N > 3 pages that are no longer linked from one to another. Hate to impute motives but after I found that I "accidentally" bought the same product multiple times because I simply couldn't trivially figure out WHAT I'd bought and hadn't yet configured I realized it was probably a revenue maximization exercise. Still use them because I like their service, but very dark.


As a long time lover of Dyn and long time hater of Oracle, this deal just makes me sick.

Anyone have good experience w/ NS1 or Neustar? I need good GSLB for active/active...


Highly recommend AWS Route53. You don't have to be using AWS to use Route53. It is solid, global anycast, and supports heath checks and geo DNS.


I'm moving my personal domains over to Route53. DYN's standard service is pretty expensive as it is.


Seconded. I still keep a DNSMadeEasy account around for some things, but Route53 just is great.


I use ns1 and am very pleased with the service. Many options for handling your responses (based on uptime, geo, etc)


Ironically they're now down - https://news.ycombinator.com/item?id=13019858


2016 keeps on delivering.


Who are they going to sue now?


The internet.

On a more serious note, does Dyn perhaps have any interesting patents?


Expect more of this stuff ahead.

These big companies have so much money sitting around doing nothing.


Any idea what happens to those of us that have legacy lifetime accounts for supporting Dyn in the early days?


If Oracle rebrands the tech they'll likely tell you the DYN products are being abandoned and you need to start paying for the new oracle branded stuff.

Time to switch if you care about your service.


It's not your lifetime it's theirs and Dyn just got put into hospice.


How long until I get screwed on my old lifetime custom zone


If you have domains.google as your registrar, they include google's dns, which has been very nice, it does support dynamic usage as well.


If you're looking for another provider I recommend http://dnsmadeeasy.com/ . It's cheaper, faster, better uptime and supports ANAME records.


Second that. Excellent uptime and pricing. Can't recommend them enough.


I'm curious - does anyone know if Dyn uses Oracle databases on the backend? Oracle seems to making a land grab for any *aaS provider that's using their DBs. Presumably because they realize their DB business is quickly being eaten by cloud providers and people sick of their "audits" (read: extortion tactics).


They tend to buy what their customers use, not necessarily their own clients. They probably identified DYN as a common provider among a subset of their customers and thought it would make for a good acquisition.


Their job postings ask for this:

"Knowledge of databases such as PostgreSQL, MySQL, Cassandra, RDS, DynamoDB"


Well I know Comodo does. I know this because if you enter any invalid data at a purchasing screen, you get Oracle error codes presented right on your screen.

Last tested a year ago.


The timimg of the acquisition with Dyn's recent DDOS has me feeling even more paranoid than usual.


That's a great DNS service you've got there; be a shame if someone were to DOS it.


"Huh, looks like the world really depends on this service. Probably means we can charge $oracle_moneys for it. Better send them an offer."


Oracle will probably end up just milking the Dyn installed base - and also try to leverage Dyn for the multiple Oracle public/private cloud efforts. However this is a great day for Dyn competitors like NS1!!


There's still a great value hiding in enterprise software for startups. De-oracling the fortune 500 or de-consulting the enterprise from enterprise vendors still looks away. But SMB SaaS growing up and eating internal departments is still possible. Or a juggernaut big Mac of enterprise with fixed price menu and a million solution providers is another.


Can anyone speculate on why this made sense to Oracle? Would it be to complement Oracle Cloud?


You don't understand: in order for Oracle to acquire it, it doesn't have to make sense: it just has to make money.


LOL, that was honestly my initial conclusion but I thought I was being too blunt.


Interesting play for Oracle. Dyn was in the news a lot lately due to IoT type DDoS vectors, are these thing connected? Is Dyn looking at some sort of lawsuit with regard to 'enabling'?


Exited for half a unicorn based up in New Hampshire!


I wonder how soon we'll see those "preeminent" brands jump.


I'm sure brands won't change. Dyn has too much of a market share for Oracle to rebrand dyn. Oracle has followed this same strategy with other acquisitions.


They got rid of Sun pretty quickly - and while they didn't have the market share at the time they did have a better reputation than Oracle did.


Sun is always the goto example for bad things haha. I should of said recent acquisitions.


Good reputation, but they were already seen as a sinking ship, IMO.


On the linked page they talk about the brands (Netflix, etc) that are using Dyn. Those are the ones I meant. I wonder if Oracle will be seen as a good thing or a bad thing to these companies.


Do we know for how much was the sale?

Any writeup fromdyn?


Right. So most folks here aren't following the "enterprise" space too much, but disclaimer: I work at Workday, a direct competitor to Oracle in the Business Management (ERP) software space. My background's in hosting & infrastructure.

Oracle are losing a lot of Oracle database licenses at the enterprise level and when they look around, the "cloud" is one of the main reasons why. They sell a bunch of jumped up 1990's applications which are mostly thin, crappy wrappers around a sophisticated database and their database, despite a lot of accounting and sales tactics that say otherwise, is still their main earner. You can't, unfortunately, see that anywhere, however, so I can't cite anything :-) So, enterprise customers, who are mostly the people buying their database are moving towards the "cloud" eating some of Oracle's lunch.

As a result, Oracle sees a growth opportunity in the "Cloud" arena - both on the infrastructure (AWS) and the application space (ERP). https://www.oracle.com/corporate/pressrelease/database-bench...

So they recently bought Dyn, Palerra & LogFire so you can see a pattern here. Dyn are just more "internet famous", but this is a pattern [1]

Lots and lots of talk recently from Ellison about their strategy.

http://www.networkworld.com/article/3123408/cloud-computing/... http://www.forbes.com/sites/oracle/2016/10/17/larry-ellisons...

It'll be interesting to see the duck tape and glue they use to jam all this crap together, but I think they're fighting an uphill battle with an incumbent that's years ahead of them on the product side. However, this is Enterprise world, and best product definitely does not always win. Oracle have a long history of selling successfully to Enterprises through fairly aggressive sales tactics. I'm not surprised if they see a market opportunity given how unfriendly AWS & GCE can be from an enterprise perspective.

[1] https://en.wikipedia.org/wiki/List_of_acquisitions_by_Oracle




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

Search: