Rackspace Unlikely to Withstand Amazon, Google Onslaught
blogs.barrons.com
blogs.barrons.com
Anecdotally, I would say Rackspace has the most intuitive interface and good telephone support. They've just released new performance instances which are cheaper than AWS: $0.08/hr for 2 vCPUs, 2GB RAM and plenty of SSD vs $0.12/hr for AWS 2 vCPU m1.medium.
Our local VPS has fantastic customer service and we know each other by name.
Amazon has the real truly global infrastructure (Rackspace UK is a separate entity to US, ASIA) and Route 53 Anycast is outstanding.
We'll be trialling out Google Compute Engine to see how it fares.
For now I see Amazon for global DNS infrastructure, but I can't see Rackspace going anywhere.
Seems like Gartner have previous with Rackspace: http://www.theregister.co.uk/2013/11/20/gartner_openstack_cr...
Also interesting reading: http://gigaom.com/2013/12/20/backbreaking-openstack-migratio...
Previously it was claimed that ever cheaper dedicated servers would wipe out their expensive offering. I think they're likely to retain a healthy business in 'cloud' for the same reason a nearly infinite supply of cheap dedicated server competitors didn't kill them previously - assuming they keep doing what got them this far, namely managed offerings and customer service. That's a value proposition Google, for one, will never even attempt to match.
But just think what would happen if Google or Amazon decided to go customer service friendly and charge for managing and hand holding. My feeling is that they could turn on that spigot at any time if they wanted to.
As for google, they couldn't interface with a real customer if their life depended on it. And the idea that they could suddenly build out a massive managed hosting service on a whim is not reflective of reality. Your claim is basically saying that no company in the world has a competitive advantage because anyone else could just start doing what that company is doing on a moment's notice.
Compare it to any "standard" vps offering and they're insanely expensive. Compared to the vast majority of dedicated hosters they're insanely expensive. Plus you're the only one on your cpu/memory/disk.
Anybody with a monthly AWS spend of $100 or more could have 2 dedicated servers, one in .eu, one in the US, for 20-30% less dollars, 5-10x more cpu, 32 Gig of dedicated memory or more, 100x more bandwitdth (and no surprise overuse charges, finding hosters that just "limit" you to 10 mbit is easy), everything non-shared and good support. Hetzner ES40 is EUR 40/month, 32Gig ram, 2Tb disk, unlimited bandwidth, similar offerings are available in the US (e.g. www.ovh.com.ca ).
Total spend for effectively ~ m3.2xlarge instance, one in US, one in EU : $120.
AWS spend for a comparable config : 1 month of m3.2xlarge : $650 5/10 Tb of internet facing traffic: $600/$1200
I think this is being generous to AWS, because an m3/2xlarge does not match the specifications of the dedicated server offerings.
Granted, these services are not as flexible as AWS. At the prices they offer though, is there anyone who really argues AWS is better for anything but their free tier ? As for the "extras" amazon offers : since they're all proprietary businesses, every amazon extra you use (e.g. S3) is something you're betting the business on.
And since none of the aws/google/... offerings are anywhere near what I want (just set this .jar in my homedirectory live, ditto for go binaries, or move-tag-and-push git publishing), I have to script stuff to make that happen anyway (I love docker).
Why does AWS have customers at all ?
Besides, in this region with messy peering agreement, AWS network also has the lowest ping time and the fewest hops.
AWS is effectively (almost) a transit-only network. That has pros and cons. On the plus side, it should have decent connectivity anywhere. On the minus side, if (almost) any link at all on a network is congested, that will almost certainly congest AWS traffic too. They're relatively unlikely to have long-running issues. Also : they're REALLY expensive.
Here's a calculator : http://calculator.s3.amazonaws.com/calc5.html
But the way things like this work a downhill slide (and I'm just pointing this out not claiming I have any more insight than what I've read) can be accelerated and make any business model not work for a company the size of Rackspace.
In other words it takes a certain amount of business and confidence of the market to stay in business.
Lose enough to someone or something else and the model no longer works very well. Especially when word gets out and it's no longer viewed as the thing to do "oh people don't rack their own servers anymore they use the cloud". [1] Keep in mind that switching from Rackspace is not trivial but it can be done it just takes time. [2]
[1] Think of anything where you might say "people will always need". Once things get to a tipping point the acceleration and downhill slide becomes really important to investors and employees who begin to jump ship in record numbers. After all people still use blackberrys but you realize that when the iphone started to take hold you wouldn't think "things will be ok for blackberry there will always be people who will use them".
[2] (I'm a customer and I'm happy with what they've done. I've even had emails returned by the President.) Noting though that they got rid of their very low end $16+- monthly VPS's.
FTFY
Used them at 2 different companies in various capacities, and they were never anything more than mildly incompetent in my experience. I called them the Blockbuster of datacenter.
I have worked with Rackspace since the dot com boom, they were one of the rare companies which actually grew when the bubble burst [citation needed]. It is unlikely that Rackspace will beat AWS or Google in short term but I dont think it is a dooms day either. Any large company will not make the mistake of relying on just one cloud service provider such as AWS (and I dont see why anyone would chose Google for a serious business). I have used rackspace for many reasons and spending around 20% of my infra cost on Rackspace (20% softlayer 40% Aws 20% Mixpanel).
It would be fun to have an review of how insightful a given analyst has been. It would also be good to have a proper 'full disclosure' policy.
Making a point that AWS owns 80% of the IaaS market has little to do with existing dedicated hosting provider's revenue!
"Servers" and "virtual machines" are two different things. You could easily run 10 VMs on a single server at the lower end of the VM price range. So without more precise measurements, these numbers mean nothing.
(What they say may still be true; you just can't tell from the numbers we're given here.)
Google being a perfect example. They leverage many economies of scale, but they do so through efficiencies that are non-trivial to implement. The datacenter model that google pioneered more than a decade ago is only now seeing penetration in the industry as a whole, and nobody can approach google's capabilities on the same score yet.
But while that helps google's search app it doesn't necessarily give them an edge when it comes to hosting. There are so many other factors such as customer engagement which are important in the hosting business. Amazon gets this right but google is the poster child for aspergers on a corporate scale. There's little reason to imagine that rackspace will be incapable of keeping up as technology advances. There's little reason to imagine that even digitalocean won't be able to keep up either.
Personally I wouldn't think it a forgone conclusion that google's hosting (app engine and what-have-you) will ever mature into an industry leading product. The problem with google is that everything that isn't search or backbone ISP is a side show in terms of revenue, so there is an absence of a culture capable of developing, delivering, marketing, maturing, and monetizing products at google. And definitely an absence of a culture capable of acquiring and dealing with customers.
This really is the core of it. I'm remembering how long it took AT&T after divestiture (mid 80's) to turn into a company that instead of a monopoly and technology had to actually do marketing. The upper ranks of the company down to the guys on the street never had to do that. It took a long long time.
Let me use an analogy. There are lots of different people who work in a "shop", but they use different tools and work on different things. There are wood workers who do carpentry, there are metal workers who work with metal, and so on. Even wood work can range from fine carpentry making use of a lathe to general contracting and building simple structures by nailing together pre-made pieces of lumber. While metal work can run the gamut from machining to iron mongery and forging to soldering and brazing and can include everything from a plumber on the one hand to someone who designs internal combustion engines on the other.
Fortunately we have the terminology to describe these different jobs reasonably well. It's not easy to confuse a plumber with a machinist. But that's not true when it comes to tech. We have very little terminology in that realm. It's non-trivial to explain the difference between a systems programmer, a DevOps automation guru, a mobile app dev, a web dev, etc. A lot of these jobs end up falling under the standard "software developer" heading. Even though that's about as silly as calling taxi drivers and truck drivers "automobile operators".
This problem of terminology extends into the tech industry as well. We have sufficiently developed mental models of things like restaurants to be able to understand the differences, and difficulties of changing, between sit-down, take out, delivery, and catering focused establishments. Yet when it comes to tech our brains go on vacation and suddenly the equivalent of a high-class sit-down restaurant switching to making pre-made meals to be sold in grocery stores seems trivial when it is anything but.
You get the same effect when people ruminate on how Apple should spend its huge cash reserves. With people throwing out crazy ideas of how Apple should buy into some completely different industry that would totally change the nature of the company, which is just as ridiculous as the idea they should switch to building rocket ships or submarines.
But both of those things have been around for much longer than tech so perhaps the roles are clearly defined for that reason. (You probably could find counter examples of course.)
Maybe this stems from the fact that the general public doesn't typically deal with "computer guys" [1] so they have no need to differentiate them. But they do need plumbers, carpenters and electricians and they know what they are. [2] (And those are also union jobs so that probably also makes titles necessary).
[1] How many times have you heard "he does something with computers?".
[2] Maybe this all stems from the fact that there was no central defining authority at any point that made any statement as to what the titles meant. Even in religion there are clearly defined titles, right? In tech (on the net at least) what do we have? RFC's a nice way of asking opinions.
Asking for real, my guess is RS does, but I don't know about the other 2.
On the other hand, it baffles me that providers such as Hetzner have not had a greater impact on the price charged by Rackspace and others. It's also surprising to see most of the best prices for dedicated servers coming from Europe and not the US.
Pair, Bytemark and Rackspace pick up the phone. Bytemark has the most chops by far, and is flexible to support custom stuff. If you just love *BSD, Pair of course.