Many developers are using VPS hosting to get their projects off the ground as it is more cost effective than dedicated hosting but gives you more control and flexibility than shared hosting. There are two main players in the virtualization market one being Xen now owned by Citrix and the other being Virtuozzo made by Parallels (formerly Sw-soft).

I’m not going to go into the Xen VS. Virtuzzo argument here, you can read a good comparison of the two at HostingFu. I prefer Xen as it seems that hosts that run it are a bit more technically knowledgeable and I don’t like the idea of Virtuozzo killing processes when I run out of memory (xen uses swap). You can find a good list of Xen vps providers at the Xen Wiki or at VPSChoice.

I wanted to compare two of the top Xen providers Slicehost and Linode. I currently have a $20/month VPS at both providers and here is what you get:

Slicehost
256 MB ram
10 GB Space
100 GB Transfer
$20/month

Linode
360 MB Ram
12 GB Space
200 GB Transfer
$20/month

Summary

Below is a short summary of both hosts to give you some background information.

Slicehost
SliceHost is a little bit more well known in the developer community and even hosts the Xen blog if that gives you any indication of how trusted they are in the community. They are based out of St. Louis, MO and run their own network along with using colo/transit from Datotel. They have a crisp web interface pictured below:

This web interface allows you to add new slices, resize existing slices (a painless process), rebuild a slice, reboot a slice, access a web console and reboot into a recovery partition. They also offer free DNS hosting with servers ns1/ns2 hosted at Slicehost and ns3 hosted at Softlayer in Dallas, TX.

Linode
Linode has been around for ages as well but has not received the press that Slicehost has (possibly because they recently switched from UML to Xen). One of Linode’s more famous websites that it hosts is infamous rapper Xzibit (I’m not sure if hosting gangster rappers gives you more street cred or not). They are based out of New Jersey and use multiple data centers; Hurrican Electric for California, ThePlanet for Texas, GNAX for Atlanta and NAC for New Jersey which gives them a footprint in the entire United states. Their web interface which is pictured below is amazingly informative and well put together:

Here you can add linodes, view graphs, reboot, rebuild, access a web based console, etc. They also offer free DNS hosting with a name server in each data center they host in.

Performance
I ran a standard unixbench against both of my hosts and I’ll paste the results below. Before you view them please note a number of factors can change these results including how many VPS servers are on the hosts, how busy they are, etc.

SliceHost
Dhrystone 2 using register variables 376783.7 9345690.3 248.0
Double-Precision Whetstone 83.1 2060.5 248.0
Execl Throughput 188.3 1385.8 73.6
File Copy 1024 bufsize 2000 maxblocks 2672.0 47061.0 176.1
File Copy 256 bufsize 500 maxblocks 1077.0 15932.0 147.9
File Read 4096 bufsize 8000 maxblocks 15382.0 390576.0 253.9
Pipe Throughput 111814.6 499706.2 44.7
Pipe-based Context Switching 15448.6 118041.6 76.4
Process Creation 569.3 2205.0 38.7
Shell Scripts (8 concurrent) 44.8 410.5 91.6
System Call Overhead 114433.5 421841.5 36.9
=========
FINAL SCORE 102.7

Linode
Dhrystone 2 using register variables 376783.7 38725011.7 1027.8
Double-Precision Whetstone 83.1 1992.8 239.8
Execl Throughput 188.3 3079.5 163.5
File Copy 1024 bufsize 2000 maxblocks 2672.0 132814.0 497.1
File Copy 256 bufsize 500 maxblocks 1077.0 37258.0 345.9
File Read 4096 bufsize 8000 maxblocks 15382.0 1219939.0 793.1
Pipe-based Context Switching 15448.6 361016.5 233.7
Pipe Throughput 111814.6 2418718.5 216.3
Process Creation 569.3 4223.0 74.2
Shell Scripts (8 concurrent) 44.8 965.7 215.6
System Call Overhead 114433.5 1959239.5 171.2
=========
FINAL SCORE 277.8

The real performance difference we see here is because Slicehost uses opterons with lower cache while Linode uses higher end Xeons.

Uptime/Reliability
Both hosts have been extremely reliabled and I haven’t noticed any downtime on either node. I currently use Pingdom to monitor both nodes on five minute intervals.

Speed
Both nodes have great throughput for transit:

Slicehost
[ross@sparta ~]$ wget http://cachefly.cachefly.net/10mb.test -O /dev/null
19:14:01 (25.8 MB/s) - `/dev/null’ saved [10485760/10485760]
[ross@sparta ~]$

Linode (ThePlanet - Dallas, TX)
[ross@rome ~]$ wget http://cachefly.cachefly.net/10mb.test -O /dev/null
19:13:45 (59.8 MB/s) - `/dev/null’ saved [10485760/10485760]
[ross@rome ~]$

Support
I don’t have much to say here as I don’t need support. However the few times I have sent questions in on both sides they have been answered very quickly.

Community
Both companies have strong user communities and active IRC channels. Slicehost has a great articles repository filled with detailed tutorials on things like nginx, ruby, etc. Linode gravitates to using the Wiki and Forum to help users or provide tutorials. I’ve been in both companies IRC channels for a bit and it seems that Linode users are a bit more helpful to people and keep a more polite tone.

Pros and Cons

Slicehost
Pro - One thing that really makes Slicehost stand out is their ability to backup your entire VPS (for an extra charge). You get three different restore points and you can schedule backups. I have this feature enabled on my account with backups running daily and weekly.
Pro - Raid 10, although I don’t see the performance increase
Con - DNS manager is difficult to use and doesn’t offer some of the great features Linode does
Con - No ability to add extra resources like ram, disk, etc. without upgrading your entire slice.

Linode
Pro - The ability in graphs to see network, cpu and disk usage real time.
Pro - A great dns manager that allows you to slave domains off of your own name servers and import records.
Pro - The ability to add extra resources without upgrading to a higher package.
Pro - The ability to add users to your account and have access restrictions. You can set it up so extra users can add linodes, manage dns, pay bills, etc.
Pro - Multiple locations so you can setup DR in geographically diverse locations
Pro - Private ips and failover
Con - No scheduled backups of your entire linode.

Conclusion
Both companies have very technical people running them and have a great track record for reliability. If you are looking for more bang for your buck then I would consider checking out Linode as they give you a bit more resources and a few more features that can be very valuable. If I had to chose between the two I would go with Linode as well as I feel their people are a bit more friendly, their features more robust and the geographic diversity they offer invaluable.

Note:
Linode owner Chris provided this article on raid 1 vs raid 10 performance and it is pretty valid as I listed raid 10 as a pro for slicehost.

Over at WHT a thread is raging over LayeredTech’s price increase on older servers. This looks to be an increase of about $22-30/month per server after an increase of $8/month not long ago. The justification is these older servers are taking up more power/space then newer servers. I have posted a few things on the thread that I will shared with you here.

First and foremost this was communicated very poorly to customers and has alienated many of them. Sending out a mass noticed like this of course was going to cause a huge customer backlash and bad press. What they should have done is call these customers up individually and try to get them to upgrade at a discount. This will allow an increase of profits and allow you to put customers on newer servers that use less power/space. Ultimately some customers would leave but I’m guessing this was their goal in the first place (get rid of lower end customers).

Next it seems many people in the thread were appalled that their server costs would ever be increased. People need to face the reality that as power/space/transit/staff/business costs go up their providers will have to pass these costs along to customers. The price of everything you pay is for is going up (gas/power/health care/bread/rice/etc.) so you should expect your hosting bill to increase in the next year as well. One of the people in the thread posted a SoftLayer promise which is:

We will never increase the prices you pay on your servers unless you upgrade a part of your server. In the event that you upgrade a server, we would only charge the difference for the part you upgrade. Now, if you order a whole new server, we would charge you at current prices.

Which is nothing but marketing propaganda. If SoftLayer’s power/transit/staff costs increase 3x over the next year they will have no choice but to pass that cost onto their customers. Making promises like that is what gets companies into trouble.

Lastly one of the problems with this price hike is the last price hike LayeredTech made promises to their customers that remain unfulfilled. If you give a person/business a good reason you are increasing their prices (e.g. remote reboot ports) they will usually be more understanding. If you don’t come through on that promise and then hit them with another price hike they will never trust you again and move their business.

In the end my sympathy goes out to every LayeredTech customer and I encourage them to look into the following providers:

Gigenet (I host with them)
TailorMadeServers (Jose is a good guy)
EliteRax (Chris and team are awesome, wish I still had a server there)
ServerCentral
Voxrox

A while back Pingdom wrote an article that compared web hosting today vs. 10 years ago (also see ReadWriteWeb’s post) and found that not much has changed. Prices have gone down and resources have increased but most shared hosting providers still rely on single server configurations with commodity control panels.

However, some developments have happened in the last two years or so. MediaTemple launched their “Grid Service” and Mosso launched their “Cloud” hosting service. These services launched with much fan-fare but haven’t really seemed to live up to their promises. MediaTemple’s Grid Service which was billed as the last hosting plan you’ll ever need has been plagued by reliability problems and Mosso’s Cloud has suffered similar issues. Both of these companies are great companies really do care about their customers in my opinion but they can’t seem to tackle the technology behind their ambitions. MediaTemple is ditching the “Grid” for their new Cluster Server and Mosso has put in place a new Chief Uptime Officer to help alleviate customer concerns.

It seems the new shared hosting platforms of today are failing to match their older counterparts in terms of reliability. You would be much better off to go with a provider like MediaLayer, CartikaHosting or Rochen if you need a stable platform for your web site. I’m not saying that these Hosting 2.0 platforms won’t improve and it looks like there may be some new platforms that launch in the near future but until then you may want to stick with what has proven to work.

I have been a customer of Pindom’s for over a year now so I thought it would be nice to write a short review on them. I signed up with Pingdom during a special where they gave everyone a year of monitoring for free and I just renewed my subscription a few months ago.

One of the great things about this company is their great interface. It seems that they put a lot of time and thought into the appearance.

Setting up your checks is a very quick and painless process and you can have different checks notify different people via email or sms.

Pingdom really separates itself from the competition in its’ reporting features. Along with standard uptime reports they also generate latency reports which is allows you to monitor your site load time.

As standard with most monitoring solutions Pingdom also offers public reports.

Things I’d like to see:
-Phone Alerts
-More Check Types (FTP, NNTP, MySQL, PgSQL, Oracle, simple port checks aren’t enough)
-Monitoring from specific locations only (I should be able to chose dallas only if I want)
-More monitoring locations (I’d like to see some pollers in Asia)

Conclusion
Overall I have to say Pingdom is one of the best monitoring solutions I have ever used especially for its price which starts at $99/yr. This solution is also one of the most accurate solutions I have ever used and I have yet to get a false positive which I routinely received with other solutions.

An interesting blog post appeared on DreamHost’s blog yesterday with Co-Founder Josh Jones posting a story about a host upgrading his VPS plan without his permission. To summarize, Josh’s VPS was going over its’ inode allotment and needed to be upgraded or it would start crashing. Instead of just letting it crash this host went proactive, upgraded Josh and notified him. Well he wasn’t happy now that he was being charged near double so he threw a fit and said he would issue a charge back, which is beyond ignorant as I’m sure DreamHost deals with a lot of charge backs which I’m sure are frustrating.

I read the entire story and I think the host was trying to be helpful, they should not have upgraded Josh without his explicit consent. However I think things could have been handled much better by Josh.

What host is Josh talking about? Well the host is Servint and the support rep was Christian Dawson. You only get the first name of the rep in the blog post but the major give away is the support person refers to the Signature VPS which matches up with Servint’s naming and pricing.

The bigger question though is why is Josh Jones Co-Founder of one of the largest shared hosting companies hosting his sites somewhere else? He had a little excerpt about testing competitors but he has had this account for three years! Maybe he just got tired of the downtime…..

We all know that outages can cost companies thousands to millions of dollars for every minute their website is down. This is of course different on whether you rely on the internet for sales or just have a web presence. For example Amazon is going to lose a lot more money if their website goes down than say a company like Exxon as the internet is Amazon’s primary sales channel.

Not only do outages hurt the sites that go down it can also greatly affect the companies hosting them. One of the things that separates a typical unmanaged host and a managed host is their SLA (Service Level Agreement). This document does what it says and defines the level or service that is expected and credits if those levels aren’t met. With a standard dedicated server you typically get %x off of your bill if you hit x amount of downtime.

However if you are a big enough player (think someone like Pepsi) and go with a managed host you can typically get a custom SLA written that could end costing the host hundreds of thousands or millions if your website goes down to a fault of the hosts. Of course one of the reasons you pay a premium for managed hosting is so you get these types of SLAs that can end up costing the host a lot of money if you go down which gives them a huge incentive to make sure you have great uptime.

This has never been more apparent then with the recent case of Internap who actually had to restate their numbers by $1.8 million because of customer credits from outages. Internap is not alone in this as SAVVIS took deep hits in the last quarter as well due to SLA credits and Rackspace had to issue some credits when it had its’ power outage.

Outages can greatly affect a managed hosting providers numbers and with that erode wall street confidence. Not only does it affect the bottom line numbers but outages can also quickly detroy a providers reputation.

After doing my financial institution host list I thought I might check out to see who hosts many of the new media websites we have today. I’m sorry if I missed your favorite web2.0 site and if I did please email me at ross [at] skafari.com and I’ll add it.

Blogs
TechCrunch - MediaTemple
Mashable - MediaTemple
ReadWiteWeb - MediaTemple
CenterNetworks - RackSpace
Go2Web20 - GoDaddy
GigaOM - SAVVIS

Social Networks
Facebook - Self (AS32934) - Transit: NTT, Level 3, Global Crossing, Hurricane Electric, Limelight
Myspace - Self (AS33739) - Transit: Level 3, Global Crossing, SAVVIS, Hurrican Electric, Internap, Limelight
Bebo - Self (AS36168) - Transit: Cogent, Abovenet, UnitedLayer
Friendster - Self (AS32338) - Transit: RCN, Internap
LinkedIN - Self (AS14413) - Transit: Internap, Limelight
Ning - Self (AS13535) - Cogent, Level 3, SAVVIS

Micro Blogging
Twitter - Verio
Pownce - Calpop
Jaiku - Nebula

Blogging
Vox / Sixapart - Typepad - Self (AS36017) - Transit: Verizon, NTT, Level 3, Global Crossing, TW Telecom, ATT
Wordpress - Peer1, Layeredtech, Rackspace, SAVVIS
Blogger - Google Property

Social News
Digg - Self (AS14833) - Transit: Cogent, Level3, Global Crossing, Shaw, Mzima, Server Cental, Limelight
Mixx - RackSpace
Newsvine - Level 3
Reddit - Servepath
Propeller - Akamai
Yahoo Buzz - Yahoo Property

Social Bookmarking
del.icio.us - Yahoo Property
Faves - Level 3
Magnolia - Servepath
StumbleUpon - Internap

Tavel
Farecast - Internap
Kayak - Akamai

Photos
Flickr - Yahoo Property
Picasa - Google Property
Riya - Ethr.net
Smugmug - Self (AS11266) - Transit: PCCW, Global Crossing, TW Telecom, Hurrican Electric, Shaw, Limelight
PhotoBucket - Self (AS14173) - Transit: Cogent, Qwest, Level 3, SAVVIS,

Audio
Odeo - Cogent
Virb - TowardEX
Pandora - Internap
Last.fm - RapidSwitch
Jango - SAVVIS
Slacker - American Digital Network

Video
MotionBox - Internap
Veoh - Self (AS40415) - Transit: Level3, Global Crossing
Vimeo - RackSpace
Blip.TV - Self (AS18559) - Transit: Cogent, Datagram
Revver - Bandcon
Viddler - Limelight

Steaming
Justin.TV - Peak Web Hosting
Usteam - Internap

Instant Messaing
Meebo - Self (AS19527) - Transit: Level 3, Internap
eBuddy - Level 3
imeem - Self (AS36119) - Transit: Cogent, Level 3, Global Crossing, Hurricane Electric, Unitedlayer, Limelight

The first thing I noticed is it seems MediaTemple hosts a lot of web 2.0 blogs but it doesn’t seem to be paying off in sales to the web 2.0 world. Also to note it seems that social networking is where the money is because most of those players have their own networks / infastructure.

I always find it interesting to see where certain industries host at or buy their transit from so I thought it would be interesting to look at the financial sector. Please note that this is just who is hosting each institution’s website and doesn’t include who they may be buying other services from.

Banks
ING Direct - Sprint
HSBC Direct - Self (AS26381) - Trransit - Verizon, ATT, Internap
CapitalOne - Self (AS19769) - Transit: Verizon, ATT
Chase - Self (AS13463) - Transit: Verizon, Internap
USBank - Self (AS3147) - Transit: Verizon, ATT
Bank of America - Self (AS10794) - Transit: Qwest, Verizon, Sprint, Telsta, Level3, Colt - Note: Prolexic ddos filtering
Washington Mutual (WaMu) - Self (AS15195) - Transit: ATT
PNC Bank - Self (AS10995) - Transit: Verizon, ATT
Regions Bank - Self (AS10801) - Transit: ATT, ITC Delta
UBS - Self (AS8883) - Transit: Verizon, Sunrise
WellsFargo - Self (AS10837) - Transit: Qwest, ATT
Wachovia - Self (AS4243) - Transit: ATT, Verizon, Sprint
Citibank - Self (AS2911) - Transit: Sprint, SAVVIS, ATT
Fifth Third - Self (AS36050) - Transit: Sprint, Cincinnati Bell
National City - Self (AS15005) - Transit: ATT, Level3, Verizon
KeyBank - Self (AS11286) - Transit: ATT, Verizon

Online Broker
ShareBuilder - SAVVIS
OptionsXpress - Self (AS32405) - Transit: Abovenet, Hurrican Electric, Internap, ServerCentral
ThinkorSwim - Self (AS26015) - Transit: Cogent, Qwest
TradeKing - Akamai
Etrade - Self (AS6352) - Transit: Qwest, Verizon, Sprint, Level3, Global Crossing, SAVVIS, TimeWarner Telecom, ATT, Internap
Ameritrade - Self (AS12192) - Transit: Qwest, Sprint, NTT, ATT, Global Crossing
Scottrade - Self (AS12221) - Transit: Qwest, Verizon, ATT, SAVVIS
Zecco - Rackspace

Mutual Funds / 401k / IRA
T. Rowe Price - Self (AS6566) - Transit: ATT, Verizon, Level3
Vanguard - ATT
Fidelity - Self (AS11040) - Transit: Verizon, Sprint, ATT
Schwab - Self (AS6949) - Transit: Qwest, Verizon, ATT
Janus - TW Telecom
AmericanCentury - Self (AS26887) - Transit: Qwest, ATT

Investments
MorganStanley - Akamai
MerrillLynch - Self (AS10998) - Transit: ATT, Verizon
EdwardJones - Self (AS14579) - Transit: Qwest, ATT, Verizon
JohnHancock - CGI - Transit:
RaymondJames - Self (AS27283) - Transit: Qwest, ATT
FranklinTempleton - Self (AS26685) - Transit: ATT
BlackRock - Qwest
Blacstone - Verizon
Lazard - TW Telecom
Ameriprise - IBM
Goldman Sachs - Self (AS6195) - Transit: Sprint, SAVVIS, ATT
Lehman Brothers - Self (AS6303) - Transit: Cogent, Verizon, Sprint, KDDI, PCCW, ATT, Reach, Internap, Global Crossing

Financial Websites
Morningstar - Self (AS11227) - Transit: ATT, Internap
Forbes - ATT
BusinessWeek - Self (AS4546) - Transit: ATT, Internap
Bloomberg - Self (AS8188) - Transit: Verizon, ATT
MarketWatch - Self (AS7105) - Transit: Sprint, TW Telecom, ATT, IPNetworks
MotleyFool - Self (AS10240) - Transit: Paetec, Internap, NTT
Bankrate - Akamai
TheStreet - SAVVIS
Marketocracy - HostedSolutions
TheUpDown - Peer1
CakeFinancial - RackSpace
Covestor - Hosting.com

You’ll notice that two names really dominate this list Verizon and ATT. It will be interesting to see if the downturn in the financial markets will affect these institutions spending and trickle down to their providers.

This also brings up the question of how much these financial institutions are outsourcing and how well companies like SAVVIS / ATT / Internap are doing when trying to sell them services.

It isn’t a secret that I’m against companies calling their product “Grid” unless they actually meet the academic standards of grid computing. This is something I have discussed in-depth with Michael Sheehan of GoGrid and we have basically agreed to disagree.

While Michael and I still disagree on terminology I still decided to give their product a shot. GoGrid allows users to deploy servers (web/database) and load balancers at the click of a button. As a current Amazon EC2 user I have to say GoGrid pleasantly surprised me.

I believe the biggest advantage GoGrid has right off is that it actually has an administrative console where you can deploy/manage servers. This is different than EC2 where you have to use third party tools/software such as ElasticFox or RightScale.

The next thing that caught my eye was the ability to do windows servers which you can see an image built in the screen shot above which is something EC2 currently doesn’t offer.

My personal favorite feature is the hardware load balancing using an F5 load balancer. This is perfect for start ups that need to deploy an application that needs high availability/high traffic. Setting this up was extremely simple and I had two servers load balanced within 5 minutes.

EC2 and GoGrid share a few common features such as persistent ips, persistent storage (linux only), and private networks. It appears that Amazon only offers an SLA for its S3 product (please correct me if I am wrong) where with GoGrid you get a %10,000 SLA However with EC2 you can utilize Amazon infinite storage platform S3 and not be billed any traffic between the two. Also, EC2 allows you to create your own images to deploy, which can save a lot of time if you are managing large clusters.

One major difference is the price in these two products. At GoGrid you are charged $.19/ram hour and at EC2 you are charged per instance hour, depending on what instance you use. If you use the small instance you are charged $.10/hr for a 1.2 ghz opteron, 1.7 GB ram, and 160 GB storage. This would be comparable to a 2 GB server at GoGrid which gives you a Xeon 3.0 ghz and 30 GB storage at $.38/hr. At near this cost you can get a large instance with 4 cores, 7.5 GB Ram and 850 GB storage with EC2.

The other thing you have to account for in the price is data transfer. Amazon EC2 charges you $.10/GB of transfer in and out where GoGrid gives you free incoming but $.50/GB outgoing. This is quite a huge difference especially if you are transferring terabytes of data. You can get the pricing down to $.17/GB if you purchase one of their pre-paid plans but this is still a pretty large difference.

Some of the features that I would like to see at GoGrid include:
- Storage Pools
- Per server bandwidth graphs
- Server monitoring (uptime/cpu/disk/etc.)
- Custom server images
- Larger images (more processor/ram)
- Hardware firewalls

Conclusion
GoGrid is definitely a product worth keeping in mind if you need an on-demand utility computing platform. It offers server key advantages over EC2 but also lacks in so many other ways so I can’t say one product is better than another.

Final thought
Please note that GoGrid is still in beta and during my testing I have noticed a few bugs.

Lately in the web hosting space industry the buzz has been going to the content delivery networks.

Some of the providers in this space are:

Akamai
BitGravity
CacheFly
Cachelogic
CDNetworks
Chinacache
EdgeCast Networks
GS Neotech
Internap
Level 3
Limelight Networks
Mirror Image
Navisite
Panther Express
SyncCast
VeriSign
VelociX

We all use these cdn networks every day without realizing it typically, some examples of this are Coke Rewards using Akami, Revision3 using Bitgravity and BlipTV using EdgeCast.

I have been wanting to do a real world speed test from multiple sites to many of these providers but my many communications have been unanswered or mishandled by many of these providers so I decided to judge the three that have public speed test files listed: CacheFly, Bitgravity, Limelight (late entrance) and VoxCDN (Voxel).

So you don’t have to go through all the raw data I’ll tell you the winner is………….. CacheFly!

Raw data below

From St. Louis, MO (67.207.143.201) [Host www.slicehost.com]

[sparta]$ wget http://speedtest.voxcdn.com/test_tenmb.data
–23:31:46– http://speedtest.voxcdn.com/test_tenmb.data
Resolving speedtest.voxcdn.com… 208.122.28.2, 208.122.28.3, 208.122.28.18
Connecting to speedtest.voxcdn.com|208.122.28.2|:80… connected.
HTTP request sent, awaiting response… 200 OK
Length: 10485760 (10M) [text/plain]
Saving to: `test_tenmb.data’

100%[=========================================>] 10,485,760 9.20M/s in 1.1s

23:31:48 (9.20 MB/s) - `test_tenmb.data’ saved [10485760/10485760]

[sparta]$ wget http://cachefly.cachefly.net/10mb.test
–23:32:37– http://cachefly.cachefly.net/10mb.test
Resolving cachefly.cachefly.net… 205.234.175.175
Connecting to cachefly.cachefly.net|205.234.175.175|:80… connected.
HTTP request sent, awaiting response… 200 OK
Length: 10485760 (10M) [application/octet-stream]
Saving to: `10mb.test.1′

100%[====================================>] 10,485,760 14.5M/s in 0.7s

23:32:38 (14.5 MB/s) - `10mb.test.1′ saved [10485760/10485760]

[sparta]$ wget http://bitcast-a.bitgravity.com/bitgravity/10MB
–23:39:51– http://bitcast-a.bitgravity.com/bitgravity/10MB
Resolving bitcast-a.bitgravity.com… 208.67.237.237
Connecting to bitcast-a.bitgravity.com|208.67.237.237|:80… connected.
HTTP request sent, awaiting response… 200 OK
Length: 10240000 (9.8M) [application/octet-stream]
Saving to: `10MB’

100%[=====================================>] 10,240,000 743K/s in 15s
[sparta]$ wget http://test.vo.llnwd.net/o1/sales/tenmegfile.wmv
–02:13:48– http://test.vo.llnwd.net/o1/sales/tenmegfile.wmv
Resolving test.vo.llnwd.net… 68.142.73.29, 208.111.156.147
Connecting to test.vo.llnwd.net|68.142.73.29|:80… connected.
HTTP request sent, awaiting response… 200 OK
Length: 11049222 (11M) [video/x-ms-wmv]
Saving to: `tenmegfile.wmv’

100%[=============================================>] 11,049,222 2.82M/s in 4.3s

02:13:52 (2.43 MB/s) - `tenmegfile.wmv’ saved [11049222/11049222]

[sparta]$

From Los Angeles, CA (72.34.62.25) [host www.medialayer.net]

[athena]# wget -O /dev/null http://speedtest.voxcdn.com/test_tenmb.data
–23:42:05– http://speedtest.voxcdn.com/test_tenmb.data
Resolving speedtest.voxcdn.com… 208.122.62.226
Connecting to speedtest.voxcdn.com|208.122.62.226|:80… connected.
HTTP request sent, awaiting response… 200 OK
Length: 10485760 (10M) [text/plain]
Saving to: `/dev/null’

100%[=======================================>] 10,485,760 1.08M/s in 9.8s

23:42:15 (1.02 MB/s) - `/dev/null’ saved [10485760/10485760]

[athena]# wget -O /dev/null http://cachefly.cachefly.net/10mb.test
–23:42:15– http://cachefly.cachefly.net/10mb.test
Resolving cachefly.cachefly.net… 205.234.175.175
Connecting to cachefly.cachefly.net|205.234.175.175|:80… connected.
HTTP request sent, awaiting response… 200 OK
Length: 10485760 (10M) [application/octet-stream]
Saving to: `/dev/null’

100%[=======================================>] 10,485,760 5.53M/s in 1.8s

23:42:17 (5.53 MB/s) - `/dev/null’ saved [10485760/10485760]

[athena]# wget -O /dev/null http://bitcast-a.bitgravity.com/bitgravity/10MB
–23:42:17– http://bitcast-a.bitgravity.com/bitgravity/10MB
Resolving bitcast-a.bitgravity.com… 208.67.237.237
Connecting to bitcast-a.bitgravity.com|208.67.237.237|:80… connected.
HTTP request sent, awaiting response… 200 OK
Length: 10240000 (9.8M) [application/octet-stream]
Saving to: `/dev/null’

100%[=======================================>] 10,240,000 2.52M/s in 4.0s

23:42:21 (2.41 MB/s) - `/dev/null’ saved [10240000/10240000]

[athena]# wget http://test.vo.llnwd.net/o1/sales/tenmegfile.wmv -O /dev/null
–02:13:39– http://test.vo.llnwd.net/o1/sales/tenmegfile.wmv
Resolving test.vo.llnwd.net… 208.111.144.221, 208.111.144.77
Connecting to test.vo.llnwd.net|208.111.144.221|:80… connected.
HTTP request sent, awaiting response… 200 OK
Length: 11049222 (11M) [video/x-ms-wmv]
Saving to: `/dev/null’

100%[=======================================>] 11,049,222 6.34M/s in 1.7s

02:13:41 (6.34 MB/s) - `/dev/null’ saved [11049222/11049222]

[athena]#

From Dallas, TX (75.126.164.186) [Host www.softlayer.com]

[ink]# wget -O /dev/null http://speedtest.voxcdn.com/test_tenmb.data
–23:37:18– http://speedtest.voxcdn.com/test_tenmb.data
Resolving speedtest.voxcdn.com… 208.122.28.2, 208.122.28.18, 208.122.28.3
Connecting to speedtest.voxcdn.com|208.122.28.2|:80… connected.
HTTP request sent, awaiting response… 200 OK
Length: 10485760 (10M) [text/plain]
Saving to: `/dev/null’

100%[==============================>] 10,485,760 363K/s in 29s

23:37:47 (351 KB/s) - `/dev/null’ saved [10485760/10485760]

[ink]# wget -O /dev/null http://cachefly.cachefly.net/10mb.test
–23:37:47– http://cachefly.cachefly.net/10mb.test
Resolving cachefly.cachefly.net… 205.234.175.175
Connecting to cachefly.cachefly.net|205.234.175.175|:80… connected.
HTTP request sent, awaiting response… 200 OK
Length: 10485760 (10M) [application/octet-stream]
Saving to: `/dev/null’

100%[=======================================>] 10,485,760 10.2M/s in 1.0s

23:37:48 (10.2 MB/s) - `/dev/null’ saved [10485760/10485760]

[ink]# wget -O /dev/null http://bitcast-a.bitgravity.com/bitgravity/10MB
–23:37:48– http://bitcast-a.bitgravity.com/bitgravity/10MB
Resolving bitcast-a.bitgravity.com… 208.67.237.237
Connecting to bitcast-a.bitgravity.com|208.67.237.237|:80… connected.
HTTP request sent, awaiting response… 200 OK
Length: 10240000 (9.8M) [application/octet-stream]
Saving to: `/dev/null’

100%[=====================================>] 10,240,000 5.98M/s in 1.6s

23:37:50 (5.98 MB/s) - `/dev/null’ saved [10240000/10240000]

[ink]# wget http://test.vo.llnwd.net/o1/sales/tenmegfile.wmv -O /dev/null
–02:09:00– http://test.vo.llnwd.net/o1/sales/tenmegfile.wmv
Resolving test.vo.llnwd.net… 68.142.121.133, 208.111.159.22
Connecting to test.vo.llnwd.net|68.142.121.133|:80… connected.
HTTP request sent, awaiting response… 200 OK
Length: 11049222 (11M) [video/x-ms-wmv]
Saving to: `/dev/null’

100%[=========================================>] 11,049,222 2.24M/s in 4.7s

02:09:05 (2.24 MB/s) - `/dev/null’ saved [11049222/11049222]

From Dallas, TX (72.249.28.154) [Host www.colo4dallas.com]

[ns1]# wget -O /dev/null http://speedtest.voxcdn.com/test_tenmb.data
–23:48:20– http://speedtest.voxcdn.com/test_tenmb.data
=> `/dev/null’
Resolving speedtest.voxcdn.com… 208.122.28.3, 208.122.28.18, 208.122.28.2
Connecting to speedtest.voxcdn.com|208.122.28.3|:80… connected.
HTTP request sent, awaiting response… 200 OK
Length: 10,485,760 (10M) [text/plain]

100%[====================================>] 10,485,760 4.63M/s

23:48:22 (4.62 MB/s) - `/dev/null’ saved [10485760/10485760]

[ns1]# wget -O /dev/null http://cachefly.cachefly.net/10mb.test
–23:48:22– http://cachefly.cachefly.net/10mb.test
=> `/dev/null’
Resolving cachefly.cachefly.net… 205.234.175.175
Connecting to cachefly.cachefly.net|205.234.175.175|:80… connected.
HTTP request sent, awaiting response… 200 OK
Length: 10,485,760 (10M) [application/octet-stream]

100%[====================================>] 10,485,760 27.07M/s

23:48:23 (27.06 MB/s) - `/dev/null’ saved [10485760/10485760]

[ns1]# wget -O /dev/null http://bitcast-a.bitgravity.com/bitgravity/10MB
–23:48:23– http://bitcast-a.bitgravity.com/bitgravity/10MB
=> `/dev/null’
Resolving bitcast-a.bitgravity.com… 208.67.237.237
Connecting to bitcast-a.bitgravity.com|208.67.237.237|:80… connected.
HTTP request sent, awaiting response… 200 OK
Length: 10,240,000 (9.8M) [application/octet-stream]

100%[====================================>] 10,240,000 7.51M/s

23:48:24 (7.49 MB/s) - `/dev/null’ saved [10240000/10240000]

[ns1]# wget http://test.vo.llnwd.net/o1/sales/tenmegfile.wmv -O /dev/null
–02:17:33– http://test.vo.llnwd.net/o1/sales/tenmegfile.wmv
=> `/dev/null’
Resolving test.vo.llnwd.net… 68.142.121.133, 208.111.159.22
Connecting to test.vo.llnwd.net|68.142.121.133|:80… connected.
HTTP request sent, awaiting response… 200 OK
Length: 11,049,222 (11M) [video/x-ms-wmv]

100%[====================================>] 11,049,222 37.91M/s

02:17:34 (37.90 MB/s) - `/dev/null’ saved [11049222/11049222]

[ns1]#

From London, UK (85.13.224.210) [Host www.coreix.net]

wget http://speedtest.voxcdn.com/test_tenmb.data
–05:12:28– http://speedtest.voxcdn.com/test_tenmb.data
=> `test_tenmb.data’
Resolving speedtest.voxcdn.com… 208.122.31.11, 208.122.31.10, 208.122.31.9
Connecting to speedtest.voxcdn.com|208.122.31.11|:80… connected.
HTTP request sent, awaiting response… 200 OK
Length: 10,485,760 (10M) [text/plain]

100%[====================================>] 10,485,760 9.62M/s

05:12:29 (9.61 MB/s) - `test_tenmb.data’ saved [10485760/10485760]

wget http://cachefly.cachefly.net/10mb.test
–05:13:43– http://cachefly.cachefly.net/10mb.test
=> `10mb.test’
Resolving cachefly.cachefly.net… 205.234.175.175
Connecting to cachefly.cachefly.net|205.234.175.175|:80… connected.
HTTP request sent, awaiting response… 200 OK
Length: 10,485,760 (10M) [application/octet-stream]

100%[====================================>] 10,485,760 11.24M/s

05:13:44 (11.23 MB/s) - `10mb.test’ saved [10485760/10485760]

wget http://bitcast-a.bitgravity.com/bitgravity/10MB
–05:14:41– http://bitcast-a.bitgravity.com/bitgravity/10MB
=> `10MB’
Resolving bitcast-a.bitgravity.com… 208.67.237.237
Connecting to bitcast-a.bitgravity.com|208.67.237.237|:80… connected.
HTTP request sent, awaiting response… 200 OK
Length: 10,240,000 (9.8M) [application/octet-stream]

100%[====================================>] 10,240,000 9.89M/s

05:14:42 (9.89 MB/s) - `10MB’ saved [10240000/10240000]

wget http://test.vo.llnwd.net/o1/sales/tenmegfile.wmv
–07:28:01– http://test.vo.llnwd.net/o1/sales/tenmegfile.wmv
=> `tenmegfile.wmv’
Resolving test.vo.llnwd.net… 87.248.212.26, 87.248.211.193
Connecting to test.vo.llnwd.net|87.248.212.26|:80… connected.
HTTP request sent, awaiting response… 200 OK
Length: 11,049,222 (11M) [video/x-ms-wmv]

100%[====================================>] 11,049,222 10.83M/s

07:28:02 (10.83 MB/s) - `tenmegfile.wmv’ saved [11049222/11049222]

From Amsterdam, NL [Host www.solidhost.com]

[office]# wget http://speedtest.voxcdn.com/test_tenmb.data
–07:22:20– http://speedtest.voxcdn.com/test_tenmb.data
=> `test_tenmb.data’
Resolving speedtest.voxcdn.com… 208.122.31.10, 208.122.31.11, 208.122.31.9
Connecting to speedtest.voxcdn.com|208.122.31.10|:80… connected.
HTTP request sent, awaiting response… 200 OK
Length: 10,485,760 (10M) [text/plain]

100%[==========================================>] 10,485,760 11.02M/s

07:22:21 (11.02 MB/s) - `test_tenmb.data’ saved [10485760/10485760]
[office]# wget http://cachefly.cachefly.net/10mb.test
–07:22:29– http://cachefly.cachefly.net/10mb.test
=> `10mb.test’
Resolving cachefly.cachefly.net… 205.234.175.175
Connecting to cachefly.cachefly.net|205.234.175.175|:80… connected.
HTTP request sent, awaiting response… 200 OK
Length: 10,485,760 (10M) [application/octet-stream]

100%[============================================>] 10,485,760 11.02M/s

07:22:30 (11.01 MB/s) - `10mb.test’ saved [10485760/10485760]
[office]# wget http://bitcast-a.bitgravity.com/bitgravity/10MB
–07:22:36– http://bitcast-a.bitgravity.com/bitgravity/10MB
=> `10MB’
Resolving bitcast-a.bitgravity.com… 208.67.237.237
Connecting to bitcast-a.bitgravity.com|208.67.237.237|:80… connected.
HTTP request sent, awaiting response… 200 OK
Length: 10,240,000 (9.8M) [application/octet-stream]

100%[==========================================>] 10,240,000 11.02M/s
[office]# wget http://test.vo.llnwd.net/o1/sales/tenmegfile.wmv
–08:37:26– http://test.vo.llnwd.net/o1/sales/tenmegfile.wmv
=> `tenmegfile.wmv’
Resolving test.vo.llnwd.net… 87.248.211.193, 87.248.212.26
Connecting to test.vo.llnwd.net|87.248.211.193|:80… connected.
HTTP request sent, awaiting response… 200 OK
Length: 11,049,222 (11M) [video/x-ms-wmv]

100%[==============================================================>] 11,049,222 8.60M/s

08:37:29 (8.59 MB/s) - `tenmegfile.wmv’ saved [11049222/11049222]

07:22:41 (11.02M/s) - `10MB’ saved [10240000/10240000]

From Chicago, IL [Host www.gigenet.com]

[avalon]$ wget http://cachefly.cachefly.net/10mb.test
–22:33:26– http://cachefly.cachefly.net/10mb.test
=> `10mb.test.9′
Resolving cachefly.cachefly.net… 205.234.175.175
Connecting to cachefly.cachefly.net|205.234.175.175|:80… connected.
HTTP request sent, awaiting response… 200 OK
Length: 10,485,760 (10M) [application/octet-stream]

100%[========================================>] 10,485,760 8.58M/s

22:33:27 (8.58 MB/s) - `10mb.test.9′ saved [10485760/10485760]

[avalon]$ wget http://speedtest.voxcdn.com/test_tenmb.data
–22:33:39– http://speedtest.voxcdn.com/test_tenmb.data
=> `test_tenmb.data.1′
Resolving speedtest.voxcdn.com… 208.122.28.2, 208.122.28.3, 208.122.28.18
Connecting to speedtest.voxcdn.com|208.122.28.2|:80… connected.
HTTP request sent, awaiting response… 200 OK
Length: 10,485,760 (10M) [text/plain]

100%[=============================================>] 10,485,760 7.59M/s

22:33:40 (7.58 MB/s) - `test_tenmb.data.1′ saved [10485760/10485760]

[avalon]$ wget http://bitcast-a.bitgravity.com/bitgravity/10MB
–22:40:04– http://bitcast-a.bitgravity.com/bitgravity/10MB
=> `10MB’
Resolving bitcast-a.bitgravity.com… 208.67.237.237
Connecting to bitcast-a.bitgravity.com|208.67.237.237|:80… connected.
HTTP request sent, awaiting response… 200 OK
Length: 10,240,000 (9.8M) [application/octet-stream]

100%[=======================================>] 10,240,000 1.08M/s ETA 00:00

22:40:12 (1.26 MB/s) - `10MB’ saved [10240000/10240000]

[avalon]$ wget http://test.vo.llnwd.net/o1/sales/tenmegfile.wmv
–01:19:06– http://test.vo.llnwd.net/o1/sales/tenmegfile.wmv
=> `tenmegfile.wmv’
Resolving test.vo.llnwd.net… 68.142.73.29, 208.111.156.147
Connecting to test.vo.llnwd.net|68.142.73.29|:80… connected.
HTTP request sent, awaiting response… 200 OK
Length: 11,049,222 (11M) [video/x-ms-wmv]

100%[===============================================>] 11,049,222 10.04M/s

01:19:07 (10.03 MB/s) - `tenmegfile.wmv’ saved [11049222/11049222]

[avalon]$

From Atlanta, GA [Host www.wvfiber.com]

[monitor]$ wget http://speedtest.voxcdn.com/test_tenmb.data
–01:56:15– http://speedtest.voxcdn.com/test_tenmb.data
=> `test_tenmb.data’
Resolving speedtest.voxcdn.com… 208.122.28.3, 208.122.28.18, 208.122.28.2
Connecting to speedtest.voxcdn.com|208.122.28.3|:80… connected.
HTTP request sent, awaiting response… 200 OK
Length: 10,485,760 (10M) [text/plain]

100%[====================>] 10,485,760 5.37M/s

01:56:17 (5.36 MB/s) - `test_tenmb.data’ saved [10485760/10485760]

[monitor]$ wget http://cachefly.cachefly.net/10mb.test
–01:56:28– http://cachefly.cachefly.net/10mb.test
=> `10mb.test’
Resolving cachefly.cachefly.net… 205.234.175.175
Connecting to cachefly.cachefly.net|205.234.175.175|:80… connected.
HTTP request sent, awaiting response… 200 OK
Length: 10,485,760 (10M) [application/octet-stream]

100%[===========================>] 10,485,760 9.81M/s

01:56:29 (9.77 MB/s) - `10mb.test’ saved [10485760/10485760]

[monitor]$ wget http://bitcast-a.bitgravity.com/bitgravity/10MB
–01:56:37– http://bitcast-a.bitgravity.com/bitgravity/10MB
=> `10MB’
Resolving bitcast-a.bitgravity.com… 208.67.237.237
Connecting to bitcast-a.bitgravity.com|208.67.237.237|:80… connected.
HTTP request sent, awaiting response… 200 OK
Length: 10,240,000 (9.8M) [application/octet-stream]

100%[=========================>] 10,240,000 347.31K/s ETA 00:00

01:57:11 (294.89 KB/s) - `10MB’ saved [10240000/10240000]
[monitor]$ wget http://test.vo.llnwd.net/o1/sales/tenmegfile.wmv
–02:17:40– http://test.vo.llnwd.net/o1/sales/tenmegfile.wmv
=> `tenmegfile.wmv’
Resolving test.vo.llnwd.net… 69.28.159.211, 208.111.161.40
Connecting to test.vo.llnwd.net|69.28.159.211|:80… connected.
HTTP request sent, awaiting response… 200 OK
Length: 11,049,222 (11M) [video/x-ms-wmv]

100%[======================================>] 11,049,222 4.45M/s

02:17:42 (4.44 MB/s) - `tenmegfile.wmv’ saved [11049222/11049222]

Update

Here are some results that people got back to me that don’t include LLNW but since one of the providers in this test is questioning it I thought I would include these as well

From Dallas, TX (66.128.59.211) [Host www.globalcon.net]

[06:10 AM] (jude:~) $ wget http://speedtest.voxcdn.com/test_tenmb.data
–06:11:14– http://speedtest.voxcdn.com/test_tenmb.data
Resolving speedtest.voxcdn.com… 208.122.28.2, 208.122.28.3, 208.122.28.18
Connecting to speedtest.voxcdn.com|208.122.28.2|:80… connected.
HTTP request sent, awaiting response… 200 OK
Length: 10485760 (10M) [text/plain]
Saving to: `test_tenmb.data’

100%[==================================================>] 10,485,760 5.67M/s in 1.8s

06:11:16 (5.67 MB/s) - `test_tenmb.data’ saved [10485760/10485760]

[06:12 AM] (jude:~) $ wget http://cachefly.cachefly.net/10mb.test
–06:13:02– http://cachefly.cachefly.net/10mb.test
Resolving cachefly.cachefly.net… 205.234.175.175
Connecting to cachefly.cachefly.net|205.234.175.175|:80… connected.
HTTP request sent, awaiting response… 200 OK
Length: 10485760 (10M) [application/octet-stream]
Saving to: `10mb.test’

100%[========================>] 10,485,760 11.2M/s in 0.9s

06:13:02 (11.2 MB/s) - `10mb.test’ saved [10485760/10485760]

[06:13 AM] (jude:~) $ wget http://bitcast-a.bitgravity.com/bitgravity/10MB
–06:14:01– http://bitcast-a.bitgravity.com/bitgravity/10MB
Resolving bitcast-a.bitgravity.com… 208.67.237.237
Connecting to bitcast-a.bitgravity.com|208.67.237.237|:80… connected.
HTTP request sent, awaiting response… 200 OK
Length: 10240000 (9.8M) [application/octet-stream]
Saving to: `10MB’

100%[==================================>] 10,240,000 10.3M/s in 0.9s

06:14:02 (10.3 MB/s) - `10MB’ saved [10240000/10240000]

From Lynnwood, WA (69.46.41.32) [Host www.gate2vn.net]

# wget http://cachefly.cachefly.net/10mb.test
–13:55:51– http://cachefly.cachefly.net/10mb.test
Resolving cachefly.cachefly.net… 205.234.175.175
Connecting to cachefly.cachefly.net|205.234.175.175|:80… connected.
HTTP request sent, awaiting response… 200 OK
Length: 10485760 (10M) [application/octet-stream]
Saving to: `10mb.test’

100%[==============================================>] 10,485,760 2.89M/s in 3.5s

13:55:55 (2.83 MB/s) - `10mb.test’ saved [10485760/10485760]

# wget http://bitcast-a.bitgravity.com/bitgravity/10MB
–13:54:19– http://bitcast-a.bitgravity.com/bitgravity/10MB
Resolving bitcast-a.bitgravity.com… 208.67.237.237
Connecting to bitcast-a.bitgravity.com|208.67.237.237|:80… connected.
HTTP request sent, awaiting response… 200 OK
Length: 10240000 (9.8M) [application/octet-stream]
Saving to: `10MB’

100%[============================================>] 10,240,000 1.42M/s in 8.4s

13:54:28 (1.16 MB/s) - `10MB’ saved [10240000/10240000]

# wget http://speedtest.voxcdn.com/test_tenmb.data
–13:55:23– http://speedtest.voxcdn.com/test_tenmb.data
Resolving speedtest.voxcdn.com… 208.122.62.226
Connecting to speedtest.voxcdn.com|208.122.62.226|:80… connected.
HTTP request sent, awaiting response… 200 OK
Length: 10485760 (10M) [text/plain]
Saving to: `test_tenmb.data’

100%[============================================>] 10,485,760 2.74M/s in 3.7s

13:55:27 (2.68 MB/s) - `test_tenmb.data’ saved [10485760/10485760]

Mosso, a provider of clustered hosting services, recently re-launched themselves. They are now calling their system “The Cloud“, which refers to cloud hosting. It seems they are jumping off the “Grid Hosting” bandwagon and using a slightly more realistic term.

Let me say I absolutely love the idea of Mosso and believe they are a true innovator of the industry. They have a platform here that allows you to do windows/linux all in one that is clustered/load balanced and easily scalable for the end user, this is absolutely amazing. Mosso also designed “The Cloud” to be reseller friendly. This takes all the administration work out of running a website or a web hosting business, you just point and click to add clients or websites.

I am not saying their platform is perfect, there have been many bumps and bruises along the way but there always are when you are the innovator. Some features still need to be added like ssh access, more reliable stats, backups that can be downloaded, etc. Overall it seems like they are pushing forward with their system and listening to their user community on features.

With the recent re-launch Mosso pared down their plans from 80 GB Space, 2000 GB Transfer to 50 GB Space, 500 GB Transfer. I think this is a bit more realistic and needed to be done to actually make a profit on customers. I would actually say they probably need to lower to the space to 20 GB and offer packages to upgrade easily.

Now comes the big mistake in my mind, they also put in a request limit to their plans which was 3 million requests (a request is each file pulled from a webpage, in a single page you can have 20-200 requests on average). I can’t stress how just absolutely amazed I was that a company on the right track could take such a huge u-turn. The problem here is that requests are not a true measure of server resource usage.

I can have a website that does extensive database queries that uses up massive resources but only uses 10 requests. Then I can have a static website that uses practically no resources generate 200 requests per page. You are unfairly punishing people just because they may have a lot of images/css/etc. files on their website. Yes, you can re-code your site (you shouldn’t have to) but if you are using Mosso as a reseller platform you can’t tell your client they need to re-code their website to your standards. These points were all fought out on the Mosso forums by myself and many other customers.

Instead of going by hits I suggested that Mosso go by actual resource usage like MediaTemple goes with their GPUs. This is an actual server resource usage metric which is fair to everyone. This was rebuffed by Jonathan at Mosso as too difficult for clients to understand. I don’t really buy this explanation as MediaTemple has tens of thousands of customers that are just fine with this measurement and understand it just fine. My guess is Mosso’s management saw this as a difficult task and is just trying to go the easy route without the understanding of what this actually does. RackSpace could have also played a part in this decision so that clients would see it as cost effective to get a server rather than keep paying to upgrade the amount of requests you get.

In the end Mosso is still not changing their stance on measuring requests; instead they have raised the plan allowance to 10 million with additional requests at costing $.03/thousand. This is a sad time for Mosso as I truly believed they were going to change the industry; however with this idiotic metric they will begin to slowly bleed their client base.

Many businesses use buzz words to attract customers to buy their products and this is especially true in the hosting industry. One of the phrases that has been catching people’s eye was “Grid” web hosting. Servepath just launched their new GoGrid product and 3tera has been screaming about their grid product for almost a year now. However, if you look at either product, you will not see a grid but rather just a cluster, which the web hosting industry has been doing for years. When you confront either company with this fact they will try to redefine the word “grid” to fit their product rather than re-engineer their product to fit the word “grid.”

As the “grid” dies down, a new buzz word is gaining ground around the industry and that term is “Green.” As people and businesses become more environmentally aware, they are going to turn to companies that offer products that can perform the same functions but have a smaller impact on mother nature. Web hosting companies are trying to market themselves as “green” in many different ways, from buying carbon credits to being totally powered by solar energy. However, there is no industry standard to rate how green a company truly is. Rackspace, Dreamhost and Thinkhost all buy the still unproven carbon credits while AISO uses solar power for their data center. Does this mean that they are all equally “green”? I would say no. While thinking about the environment is always admirable, there is a big difference between buying carbon credits and being totally powered by solar energy.

While there is no industry standard for being “green,” we can all do our part to lessen the environmental impact our industry produces in a few ways. We can begin by using servers that use less power and move on to data centers that are cooled naturally in winter. Power your data center with renewable energy such as wind, solar, or hydroelectricity. If that isn’t an option, look to your local power company to see if they have a program such as Ameren’s PurePower or DTE’s Green Currents that provides renewable energy to businesses/consumers. If we all just think a little greener not only will our companies benefit but the big winner here will be mother nature.

I wrote an article the other day about web 2.0 website hosting. In my article I discussed a number of options and a number of companies, one of those being Joyent. Well it appears Jason from Joyent wasn’t happy with my post. He commented on it telling me they didn’t have downtime and that I was doing “false reporting.”

Well it appears the Web Hosting gods have smote Jason for his insolence with a rather embarrassing downtime. The article is covered on TechCrunch one of the most popular web 2.0 blogs on the internet and was based off of an article from Datacenterknowledge.

All I can say is a three day downtime is rather embarrassing, even more so when your backups are “corrupted.” This may not be Textdrive but it seems that if you use any part of this companies services you will have problems/downtime.

Finding good web hosting for developers can be a daunting task depending on what you want to do. With all the new frameworks/applications out there it is hard for web hosts to do a one size fits all approach. Hosts are moving to meet the needs of customers who want to embrace Ruby on Rails ( RoR ), Django, Mongrel, Plone, Drupal, Svn, Trac, Turbogears, etc. We will go over the different hosting options and some hosts who may be able to fit your needs.

Shared Hosts:
Webfaction ( http://www.webfaction.com )
TextDrive / Joyent ( http://www.textdrive.com )
MediaTemple ( http://www.mediatemple.net )
Segpub ( http://www.segpub.com )
ASmallOrange ( http://www.asmallorange.com )
Dreamhost ( http://www.dreamhost.com )

VPS / Managed VPS:
Slicehost ( http://www.slicehost.com )
Engineyard ( http://www.engineyard.com )
VPSlink ( http://www.vpslink.com )

Shared Hosts

Many developers are not system administrators so a shared host that can fit their needs definitely has its advantages. It also allows the developer to focus on their website/application rather than having to worry about administrative tasks. Cost savings can also be significant for people with tight budgets working on smaller projects.

Webfaction is a web host who is really trying to work with developers. They offer many applications/frameworks right out of the box with their own custom control panel. While I do recommend WebFaction to everyone I worry about their new plans which oversell like crazy and could hurt performance in the future. Overall if you are a developer and need hosting I would go with these guys.

TextDrive was started with seed capital from customers by selling lifetime accounts and has grown to be a decent sized host. One of their key selling points is their connector which allows for better team collaboration. I have been following them for a while and my recommendation is to stay away. As I type right now the Joyent website is down and this is basically how their service goes. If you do a Google search or look on their forums you’ll notice major outages on servers, clients not being responded to in tickets and planned outages not being communicated.

MediaTemple launched their GS ( Grid Server ) hosting packages to much press and praises, since then it has been all down hill. They continue to have major outages and slow downs which is sending many of their customers to other hosts. I do have to say MediaTemple has been very up front about this and has communicated very well with their customer base. They are also planning a new service called CS ( Cluster Server ) which they say will bring much better performance/reliability to their customers. As of right now I can’t recommend MediaTemple but I hope in the future they fix the issues they are facing.

Segpub is a web host out of Australia that uses servers at Rackspace to host their clients. They are a smaller host and are quite expensive. However I have never seen a negative review of their services and all their clients boast about their reliability.

ASmallOrange is a pretty popular host that is know for their excellent customer service. I wouldn’t really say they are geared towards developers although they offer Ruby on Rails in their packages.

Dreamhost is one of the largest hosts on the planet but that doesn’t mean they do a good job. People typically flock to this host for cheap prices. They continue to have wide spread outages, slow servers and blacklisting issues. If you want your site to be up and speedy I would stay away from these guys.

Some other hosts in this arena that I don’t have a great knowledge of but you may want to look at include:
Avlux ( http://avlux.net )
PlanetArgon ( http://www.planetargon.com )
Rails Playground ( http://www.railsplayground.com )
Apis Networks ( http://apisnetworks.com )
Hosting Rails ( http://www.hostingrails.com )

Slice / VPS / Managed VPS

When your application/website outgrows shared hosting you will need to upgrade to something that can handle the load. This usually means either going with a VPS, Dedicated Server or just a company that specializes in application hosting. Many of the shared hosts above offer vps/dedicated server plans but here we are going to try and list some companies that offer these services specifically.

Slicehost has been getting much well deserved fan fare because of their hosting platform. They offer XEN vps’s with an easy to use control panel which allows you to create “slices”, backup/restore, and manage dns. This type of hosting may not be for novices as it will require you to configure your own server and install your own application. However, Slicehost does have an extensive knowledge base and community to help you in all your endeavors. I fully recommend Slicehost and even use them myself for hosting some projects!

EngineYard is one of the more well known rails hosts out there and they just closed a $3.5 million round of funding. They have a number of bigger web 2.0 clients using their service and all seem very pleased with what they get. At EngineYard you buy slices depending on your applications needs.

VPSlink developed a specific vps instance just for RoR developers which may make it easier for developers to get up and running. I tested them not to long ago and I feel indifferent towards them. Their control panel was a bit clunky and their support was just okay.

Some other slice/vps/managed service providers in this area that you may want to look at includes:
BlueBoxGrid ( http://www.blueboxgrid.com )
Thoughtbot ( http://www.thoughtbot.com )
Brightbox ( http://www.brightbox.co.uk )
Rails Machine ( http://railsmachine.com )

Dedicated / Managed

If you are a developer or have a team of people and know what you are doing you just may want a dedicated server / managed server solution. I just wanted to throw out some companies that can probably fit your needs whether you need a dedicated server or a true managed hosting provider:
Gigenet ( http://www.gigenet.com )
Voxel ( http://www.voxel.net )
Softlayer ( http://www.softlayer.com )
Handynetworks ( http://www.handynetworks.com )
Rackspace ( http://www.rackspace.com )
Servepath ( http://www.servepath.com )
Logicworks ( http://www.logicworks.com )

Conclusion

Whether you are just a small one person operation or if you are a team of people working on the next big web 2.0 site you have plenty of options. Many of these companies specialize in hosting developers and can grow with you so you can become the next Twitter, Digg, Facebook, whatever site that is popular on TechCrunch this month.

So I wanted to create a quick list of hosts and services that people can review quickly:

Shared Hosts
MediaLayer - http://www.medialayer.com
UnitedHosting - http://www.unitedhosting.com
ASmallOrange - http://www.asmallorange.com
Rochen - http://www.rochenhost.com
DowntownHost - http://www.downtownhost.com
Segpub - http://wwww.segpub.com
Fluidhosting - http://www.fluidhosting.com
Pair - http://www.pair.com
Doreo - http://www.doreo.com
Webfaction - http://www.webfaction.com
BlackKnight -http://www.blackknight.ie
Reliablesite - http://www.reliablesite.net
Dotable - http://www.dotable.com
CartikaHosting - http://www.cartikahosting.com

VPS
MyriadNetwork - http://www.myriadnetwork.com
YFVPS - http://www.yfvps.com
Zone - http://www.zone.net
Knownhost - http://www.knownhost.com
SolarVPS - http://www.solarvps.com
Slicehost - http://www.slicehost.com

Dedicated Servers
Gigenet - http://www.gigenet.com
Steadfast - http://www.steadfast.net
Voxrox - http://www.voxrox.com
CoreIX - http://www.coreix.net
Servercentral - http://www.servercentral.net

Managed/Complex
Savvis - http://www.savvis.com
Rackspace - http://www.rackspace.com
Datapipe - http://www.datapipe.net
Logicworks - http://www.logicworks.net
Internap - http://www.internap.com

Monitoring
Pingdom - http://www.pingdom.com
Hyperspin - http://www.hyperspin.com

So I thought it would be interesting to see who hosts your presidential candidate.

Democrats

Hillary Clinton
Website: www.hillaryclinton.com
Registrar: Network Solutions
Webhost: Counterpoint Networking

Barrack Obama
Website: www.barackobama.com
Registrar: GoDaddy
Web Host: Pair Networks

John Edwards
Website: www.johnedwards.com
Registrar: GoDaddy
Web Host: PlusThree

Chris Dodd
Website: www.chrisdodd.com
Registrar: GoDaddy
Web Host: DnsPark / Voxel

Joe Biden
Website: www.joebiden.com
Registrar: Network Solutions
WebHost: Truenet

Mike Gravel
Website: www.gravel2008.us
Registrar: Tucows
Web Host: MediaTemple

Dennis Kucinich
Website: www.dennis4president.com
Registrar: GoDaddy
Web Host: Hostway

Republicans

Ron Paul
Website: www.ronpaul2008.com
Registar: 1and1
Web Host: Terra Eclipse / RackSpace

Mitt Romney
Website: www.mittromney.com
Registrar: Network Solutions
Web Host: RackSpace

Rudolph Giuliani
Website: www.joinrudy2008.com
Registrar: GoDaddy
Web Host: MediaTemple

John McCain
Website: www.johnmccain.com
Registrar: GoDaddy *Private Registration
Web Host: Smart Tech Also hosts www.rnc.org and most GOP websites

Fred Thompson
Website: www.fred08.com
Registrar: Network Solutions
Web Host: HostMySite

Sam Brownback
Website: www.brownback.com
Registrar: Network Solutions
Web Host: Rackforce (Canada)

Recently there was a huge controversy around Godaddy suspending the seclists.org domain run by Fyodor (creators of things like nmap, etc.). The domain was suspended because one of its mailing list members had posted some MySpace username/passwords to the list. Apparently Godaddy called Fydor and gave him exactly one minute before they suspended his domain, you can read the story here. This caused a huge outcry in the security community who I guess was completely obvilvious to what everyone in the domain/web hosting industry has known for a long while, don’t register through Godaddy. Godaddy has been infamous for suspending domains and actually uses it as a source of revenue (they charge you money to get the domain unsuspended) and this is not the first time this “policy” has caused outcry. On January 9th Godaddy took Nectartech datacenter offline by suspending their domain because one of their customers servers had a phishing site (which goes to show their abuse team is a bunch of morons).

In response to being suspended Fyodor started a site called NoDaddy where he shares the story of what happened, alternative registrars and other things. This suspension has sent a wake up call to domain buyers who are now scrambling to find an alternative and has even spawned a CNET Article which lists registrars which are “free speech friendly”. This actually made me laugh because Fyodor and CNET both use/list registrars who are famous for suspending domains.

Nodaddy is registered at Directnic who has taken the same approach as Godaddy and has started suspending domain customers without just cause. I would say the folks over at CNET need a clue and should leave the recommendations to those of us who actually have one. So in response to all of this I’m going to give a quick rundown of recommended registrars and registrars to stay away from.

Recommended:
Moniker - Moniker is the top registrar for serious domainers (people with more then 100+ domains) and with good reason. Their interface is very clean, their prices are great and their features are plentiful. You can always reach someone by phone if you need to, I use them nearly exclusively.
Namecheap - Namecheap is an Enom reseller but they truly are good at what they do. Their interface is clean and whois privacy is included. Their prices are great and support is always friendly.
SRSPlus - SRSPlus is a subsidiary of Network Solutions. Their email support seems a bit slow but their phone support is very good. It does cost $250 to join because you will be a domain reseller but that money goes to future domain purchases.
Rebel - Rebel is part of the Momentous Corp and their interface is all web 2.0. It is very clean and pretty fast, they are lacking whois privacy right now but hopefully will include it in the future

Stay away from:
Godaddy - Godaddy as you can see is famous for suspending domains without court orders for whatever they deem necessary.
Directnic - Directnic has also taken the path of suspending domains without cause or court orders, stay away.
Dynadot - Dynadot also has horrible terms and has taken the “we will suspend domains for whatever we want” stance.
Registerfly - Registerfly is horrible and I can’t say stay away enough. Thousands have been screwed by this company.
Dotster - Dotster seems to not have a clue, their phone support wait times are ridiculous and email support is virtually non-existant.
1and1 - 1and1’s prices are low but their support is based out of the Phillipines which makes them horrible to deal with.
Gandi - Gandi’s support is some of the slowest in the industry, I recommend staying away from these guys.
Itsyourdomain - I’ve heard horror stories about these guys with renewals, I’d stay away

Neither recommended or hate:
Domainsite / Name.com - I don’t recommend these guys because their support seems a bit slow and they don’t offer privacy on their .org registrations.
4domains - 4domains management interface is extremely ugly and diffcult to use
Enom - At this moment in time enom is having very large support issues and therefor cannot be recommended.
Joker
Domainit
GKG
Planetdomain

If you have questions about a certain registrar please email me at ross [at] thehostguru.com and I will respond as I have done a fair bit of research in this market.

Okay okay I know this site is supposed to only be about web hosting but I need to go on a tanget for a minute, not because I can but because I must. I have a strong need to talk about the genius behind some of the new web sites on the internet and how it all makes me wish I just had the next billion dollar idea like the guys over at (insert next web 2.0 company fueled by Red Bull with a 19 year old at the CEO seat who will soon sell to (insert company with too much money here i.e. Google/Yahoo/Ebay/Microsoft/etc.))

During the past few years the internet has been revolutionizing into more of an interactive experience, you no longer surf the web you become a piece of it with your MySpace profile. Instead of having garage sales we sell our useless junk of Craigslist and Ebay. Instead of turning on the stereo we stream music from Pandora and Purevolume (which also recommends us music based on the songs/artists we like). Photo albums are a thing of the past with the indefinite amount of photo sharing sites like Flickr and Smugmug coming online.

Everyone talks about the internet boom of the 90’s and the subsequent burst in 00′/01′. However I don’t think that boom is anything compared to what we have seen since around 2003. Web 2.0 companies are launching daily offering everything from financial tracking to website hot spot tracking to people tracking.

Some sites as lately make me pretty angry, not because they are totally useless (they are) but because I didn’t think of them first. The perfect example of this is the Million Dollar Home Page brain child of Alex Tew. Who in their right mind would have thought some kid selling pixels on a website for a dollar a piece could actually sell a million of them. Not only that but he has launched a new site to see if lighting can strike twice.

The Million Dollar Home Page didn’t bother me too much because people actually got advertising out of the deal, I guess thing that bothers me is people buying “virtual goods” such as weapons for an online video game. My real reason for writing this is that not to long ago I found out about a site called SecondLife, apparently I’m not part of the cool crowd and hadn’t gotten the memo about it. If you haven’t heard about it yet we need to demote you to a level twenty-two paladin now, sorry.

Secondlife is a virtual world where people use REAL money to buy virtual money so they can buy virtual goods. I’m going to pause for a moment to let that seep in because it took a while for me to…………..okay. So people trade in real currency for this online currency called lindens so they can rent/buy virtual goods such as land, clothes for their character, building materials, etc.

When I heard about this I wept for the internet as we know it, not only for the internet but humanity. I remember someone on Fark saying “We need to have another great depression so we can truly value money again”, I have to agree. People actually use their limited financial resources to buy virtual beer; not only that but they go to virtual strip clubs and watch virtual strippers. Now when I say virtual strippers I don’t mean some girl is on a webcam showing you the goods I mean you see an animated figure stripping, yes people actually pay for this!!! I don’t know about the rest of you but it just wouldn’t be the same without the smell of ether mixed with alcohol that you get at the local skin dive but I guess this is a dream come true for geeks (especially the anime ones) with germaphobia.

Now it isn’t like the internet has been the first place where people have capitalized at selling totally useless goods, does anyone remember the pet rock? The internet has just brought the concept onto a whole new scale turning what use to be quick lived fads into long term businesses.

I’m not angry at the businesses I’m just waiting for my turn to come when I can make a totally useless product that sells like hotcakes to the uneducated masses (if you went to a typical American public school that means you too). Long live the useless virtual goods and the people who buy them so that young kids that play too much Halo in their parents basements can become billionaires, god bless you typical dumb american and god bless the internet that allows you to waste your money on virtual “life”.

This article is sponsored by Gigenet

Today is a day that web hosting provider Media Temple would most likely want to forget. The web hosting company launched their new “grid” server offer which promised to be “The last hosting plan you’ll ever need.” Media Temple’s offering got a huge boost by Web 2.0 site TechCrunch in this post. The ironic thing is while TechCrunch was praising the “grid” and being hosted by Media Temple they were actually on their own servers and not the grid. The TechCrunch web site stayed up during the entiire time their readers sites were down. How much downtime did “grid” server client suffer, enough that Netcraft noticed. We’ll take our own monitoring service’s stats though which said the grid has suffered 17 outages in November with 3 hours and 10 minutes downtime, you can read the ticket by Media Temple here.

Now there have been some major discussions on Media Temple’s “grid” offering a here here here and a thread about LayeredTech’s thegridlayer.com here which is powered by vendor 3tera. One thing i especially wanted to note was this thread here in which you see me post a statement to Vlad Miloushev CEO of 3tera noting that he blew through $16+ million of VC funding when he was CEO of Z-Force which crashed and eventually became attune systems (with more VC funding [they secured 14 million November 14, 2006]).

In my recent interview with Tim Dorr of A Small Orange I asked him about Media Temple’s “grid” and he responded by saying that it was not a grid but a cluster of servers and I have to agree with him on this (as would most people). The entire thought behind a grid was that you can use the computing power behind all the nodes to solve a single problem and if any one node failed there would be no impact. We’ll let IBM define a grid:

Grid computing enables the virtualization of distributed computing and data resources such as processing, network bandwidth and storage capacity to create a single system image, granting users and applications seamless access to vast IT capabilities. Just as an Internet user views a unified instance of content via the Web, a grid user essentially sees a single, large virtual computer.

You can also view more definitions of what a grid is here here here and here with example of grids. Once you read the definition of a grid and then look at Media Temple’s grid or 3tera’s Applogic you’ll notice they are not grids. They cannot distribute processing tasks across multiple nodes and if the node you are on crashes you will go down.

Why does all this matter? Because it appears that 3tera is knowingly making false claims about their product using buzz words while subtley dropping hints that they will be looking for VC funding and Media Temple is using the buzz words to make their offer more attractive to customers in the hosting market in order to grab market share. One thing is for sure in this entire mess, the winners are the marketing departments for both companies with the losers being their customers.

Since i can’t list every good host out there I thought I’d take a one post and lists some hosts that I wouldn’t recommend that you host with.

WebHostPlus / NetBunch / Dinix
Stop!!! Warning!! Don’t Do it!!! This company is infamous for it’s horrible customer service, downtimes and shady business practices. They tend to accquire companies and totally trash them (i.e. managed.com, dinix, netbunch, etc.)

Burstn