96MB Extreme 4 Hours VPS Review–Linode
It is a Friday evening and while I was browsing on the LowEndTalk and came across of this post. Instantly I thought it would be an interesting challenge for me to write an review given only a 4 hour window for me to try out the VPS (and another 2 hours to do the writing on a sunny Sunday morning). Granted there will be many shortcomings for such a review, since I might just be lucky to hit a “good” 4-hours window on the node, however I thought it would at least offer the 96MB.com readers some insight of Linode, who’s price is way beyond the standard LEB pricing, without sacrificing my own wallet so much.
So without further delay (since I only give myself two hours of writing time), let’s get started.
Basic Information and Set Up
According to this blog entry on Linode, the process for a 4 hour trial node is fairly simple. Basically all you need is a verified email address (and obviously a username and password), and you can fill up a form on their home page:
once you have the email address verified, you will be able to start a trial:
Basically you can create a VPS in any of their datacenters, and according to their payment page, their Linode 512MB plan has the following resources, note that on the top of the page, there is a small timer there informing you the total amount of time left in the trial. I believe the timer starts the moment you clicked on the “Continue with Free Trial” button, not when the VPS is actually boot up.
Also, the other interesting bit to note is that if once you have a trial VPS, if you would like to pay for it, you can upgrade to a full VPS without losing any of your configurations/data, however a reboot will be required. Linode starts the billing on the first day of the month and the first bill would be pro-rated to the first of the next month as you can see in the payment page. Finally, Linode only accept credit cards.
I thought testing the VPS in their Japan datacentre would be a good idea, simply because VPS in Asia are rare to see due to the cost and if I remembered correctly, this is probably one of the first, if not the first review that I have done for a VPS located in Asia. They have several other datacentres that you can choose from and the respective speedtest is on their website:
Once the VPS is created, you have to manually boot it up, and you can see the VPS showing on the main dashboard, with the amount of bandwidth available:
I believe for Linode, the bandwidth is pooled, which means if you have two Linode 512 and has a total of 4TB of bandwidth allowance, you can use 3TB on one server and 1TB on the other without getting charged anything extra, which I think would be neat. Obviously the graph seems to be a bit out of proportion for 1% of the graph (and it is actually more like 0.05% based on 1GB out of 2TB).
I think one of the biggest reasons why people are buying Linode is because of their control panel, which is really a piece of art. Once enters the Dashboard, the details of the VPS can be seen, as well as a log of actions performed on the VPS at a node level:
You can edit several component of the VPS configuration including the main disk. Note that the swap is included in your disk quota and you can create a swap image of up to 512MB.
By default, I chose 256MB of swap space during creation, however, if in this case I need more swap space for whatever reason, I would need to shrink my main disk first to make room for the swap space. I can also shrink my swap space and make room for the main storage disk as well. Obviously it is something easy to do at in command line, but it is pretty cool that users would be able to do it on the control panel:
The Remote Access tab shows you the IP addresses as well as any other associated information that you will need to access your VPS remotely. Note that there is only 1 IPv4 and a /64 IPv6 address given by default, however you can buy more IPv4 addresses and for IPv6, I really doubt if you’ll need that many (to get an idea of how many IPv6 addresses there are in a /64 block, feel free to read this short article here):
Instant rDNS can be assigned as well, by clicking on the Reverse DNS button. Note that if you have multiple IP addresses, you can use their IP failover and IP swap functionalities to give you a kind of HA set up.
The emergency access console Linode has is called Lish, which is an Ajax-based terminal that you can use:
The next tab is Rebuild, which as the name suggests, you can wipe out and reinstall the VPS:
Note that you have quite a few options to choose from here when it comes to selection of OS:
Arch Linux 2012.10 Arch Linux 2012.10 64bit CentOS 6.2 CentOS 6.2 64bit Debian 6 Debian 6 64bit Fedora 17 Fedora 17 64bit Gentoo Gentoo 64bit openSUSE 12.1 openSUSE 12.1 64bit Slackware 13.37 Slackware 13.37 64bit Ubuntu 10.04 LTS Ubuntu 10.04 LTS 64bit Ubuntu 11.10 Ubuntu 11.10 64bit Ubuntu 12.04 LTS Ubuntu 12.04 LTS 64bit Ubuntu 12.10 Ubuntu 12.10 64bit Arch Linux 2012.07 Arch Linux 2012.07 64bit CentOS 5.6 CentOS 5.6 64bit Debian 5.0 Debian 5.0 64bit Fedora 16 Fedora 16 64bit openSUSE 11.4 openSUSE 11.4 64bit Slackware 13.1 Slackware 13.1 64bit Ubuntu 11.04 Ubuntu 11.04 64bit Ubuntu 8.04 LTS Ubuntu 8.04 LTS 64bit
As you can see, Linode not only has the newest OS for major Linux distributions but also the older version for those who can’t use the newest version.
Other than rebuild, you can also boot into the rescue mode to save your files. You can also directly reset your root password using the control panel:
Since I was on a trial mode, in order to resize, I have to purchase an actual Linode (which I did not do), however as you can see, the process seems to be fairly simple and the Linode will be migrated automatically. They even have an estimate on how long the migration process will take.
If you have multiple Linodes, you can also clone the Linode, which would be helpful if you have, say, an HA set up with multiple VPS for the cluster. You can also clone one of the disks if you prefer.
The graphs initially did not show anything for me and I thought it was an error on their end:
Turned out they simply didn’t have sufficient data to plot the graph as the default graph shown in daily for the past 30 days. Once I click on the Last 24 Hours, the graph will start to display, as you can see the CPU, network and disk I/O information are all displayed (yes, I know I have been using quite a bit of CPU and disk I/O power, which I have been getting emails during my short trial with them, but those are necessary to perform the tests):
Settings tab will provide you the ability to change the Linode label as well as enable or disable the shutdown watchdog (I guess a sort of “autoboot” feature), but mostly importantly, it will set up the threshold for you to get alert emails so that you don’t have to keep an eye on the VPS 24×7 to know if anything bad happens:
Here is one of the automated email that I have received from them after running the DD tests:
![SNAGHTML543e1d1[5]](https://www.96mb.com/wp-content/uploads/2013/03/SNAGHTML543e1d15.png "SNAGHTML543e1d1[5]")
Finally, on the extra tab, you can purchase additional resources on the extra tab:
The prices, however, are not cheap (just like the Linode itself):
For the RAM, you are looking at 5 USD per month for each 90MB block, you can also pay yearly or two years at once:
Additional 90 MB RAM - (add $5.00/mo) Additional 90 MB RAM - (add $60.00/year) Additional 180 MB RAM - (add $10.00/mo) Additional 180 MB RAM - (add $120.00/year) Additional 270 MB RAM - (add $15.00/mo) Additional 270 MB RAM - (add $180.00/year) Additional 360 MB RAM - (add $20.00/mo) Additional 360 MB RAM - (add $240.00/year) Additional 360 MB RAM - (add $480.00/2 years)
And for disk space, each 2GB block costs an additional 2 USD per month, and you can prepay for a year’s worth of disk space as well:
Additional 2 GB Disk Space - (add $2.00/mo) Additional 2 GB Disk Space - (add $24.00/year) Additional 4 GB Disk Space - (add $4.00/mo) Additional 4 GB Disk Space - (add $48.00/year) Additional 6 GB Disk Space - (add $6.00/mo) Additional 6 GB Disk Space - (add $72.00/year) Additional 8 GB Disk Space - (add $8.00/mo) Additional 8 GB Disk Space - (add $96.00/year) Additional 10 GB Disk Space - (add $10.00/mo) Additional 10 GB Disk Space - (add $120.00/year) Additional 12 GB Disk Space - (add $12.00/mo) Additional 12 GB Disk Space - (add $144.00/year)
Bandwidth upgrade cost 10 cents per GB per month, which is why quite a few Linode users were happy when Linode announced that they will upgrade the bandwidth to the existing clients for free, by 10 time of their original bandwidth:
Additional 100GB/month Transfer - (add $10.00/mo) Additional 100GB/month Transfer - (add $120.00/year) Additional 200GB/month Transfer - (add $20.00/mo) Additional 200GB/month Transfer - (add $240.00/year) Additional 400GB/month Transfer - (add $40.00/mo) Additional 400GB/month Transfer - (add $480.00/year) Additional 600GB/month Transfer - (add $60.00/mo) Additional 600GB/month Transfer - (add $720.00/year) Additional 800GB/month Transfer - (add $80.00/mo) Additional 800GB/month Transfer - (add $960.00/year) Additional 1000GB/month Transfer - (add $100.00/mo) Additional 1000GB/month Transfer - (add $1,200.00/year) Additional 2000GB/month Transfer - (add $200.00/mo) Additional 2000GB/month Transfer - (add $2,400.00/year) Additional 5000GB/month Transfer - (add $500.00/mo) Additional 5000GB/month Transfer - (add $6,000.00/year) Additional 10000GB/month Transfer - (add $1,000.00/mo) Additional 10000GB/month Transfer - (add $12,000.00/year)
The only probably relatively reasonably priced (in low end VPS standard) add-on is the IP, which is 1 USD per month per IP:
Additional 1 IP Address - (add $1.00/mo) Additional 1 IP Address (annual) - (add $12.00/year) Additional 1 IP Address (biennial) - (add $24.00/2 years)
Other items on the main dashboard includes the ability to add Stackscripts to the VPS, which is a set of script that you could automatically add to the VPS during creation, which could be anything like changing the SSH port or allow key-only access to deploying a full software stack, which would be really useful if you need to deploy several machines or you would like the save the hassle of deploying the same software stack after you reimaged the VPS. In fact, Linode even has several stack scripts that are already written for you here:
You can also add so-called NodeBlanacers, which would cost extra and I am not sure how is that different from adding another Linode:
Linode also includes a DNS manager which people could make use of, which you can either import an existing zone for an active domain directly or create a new zone, the import zone interface is pretty intuitive and straight-forward to use:
And to add a new zone, the task is fairly easy as well. In fact, the DNS manager even has the ability to set up some of the basic records for you based on the IP address of the VPS you have with Linode:
Here is a test set up I did and as you can see, some of the default records are set up once I click the button:
You can also create a slave zone, which I think would be great as a backup DNS if you prefer to run your own DNS server (or if you worried your primary DNS provider might fail):
The Account tab is just showing the sign up form, which we will skip here, and the last tab in the control panel is support, which gives you an option to open up a support ticket when needed, although they have already had a really a really good support system (which includes the FAQ, Library, Forum, and IRC), which you should be able to find a lot of answers by yourself:
P.S.: I just noticed I have already used more than 2 hours by the time I got here, but I guess in the interests of the 96MB readers, I will just keep going. Afterall, having a complete and detailed review is definitely a lot more important to me than the little 2 hours’ challenge.
Test of the VPS
As mentioned above, the VPS that I have signed up for testing is a 512MB RAM Linode with 2TB of bandwidth and is located in their Tokyo, JP datacentre. I have installed Debian 6 32 bit OS template for testing. The VPS, as with all other VPS from Linode, uses Xen virtualization.
When the VPS was first reimaged, 25MB of RAM was used, which is a little higher than usual:
free -m
total used free shared buffers cached
Mem: 496 46 449 0 2 18
-/+ buffers/cache: 25 470
Swap: 255 0 255
Top showing the processes running:
top - 21:18:10 up 14 min, 1 user, load average: 0.00, 0.01, 0.01
Tasks: 79 total, 1 running, 78 sleeping, 0 stopped, 0 zombie
Cpu(s): 0.0%us, 0.1%sy, 0.0%ni, 99.9%id, 0.0%wa, 0.0%hi, 0.0%si, 0.0%st
Mem: 508100k total, 48048k used, 460052k free, 2496k buffers
Swap: 262140k total, 0k used, 262140k free, 18968k cached
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
1775 root 20 0 6088 2860 1536 S 0 0.6 0:00.20 bash
1772 root 20 0 8308 2824 2304 S 0 0.6 0:00.05 sshd
1661 ntp 20 0 4712 1880 1456 S 0 0.4 0:00.05 ntpd
1597 root 20 0 28492 1364 1016 S 0 0.3 0:00.00 rsyslogd
1818 root 20 0 2488 1064 816 R 0 0.2 0:00.00 top
1750 root 20 0 5544 932 544 S 0 0.2 0:00.00 sshd
1426 statd 20 0 1988 784 636 S 0 0.2 0:00.00 rpc.statd
1690 root 20 0 4396 744 580 S 0 0.1 0:00.00 cron
1087 root 16 -4 2308 728 400 S 0 0.1 0:00.06 udevd
1 root 20 0 2084 708 604 S 0 0.1 0:01.39 init
1158 root 18 -2 2304 620 296 S 0 0.1 0:00.00 udevd
1154 root 18 -2 2304 616 292 S 0 0.1 0:00.00 udevd
1414 daemon 20 0 1860 492 396 S 0 0.1 0:00.00 portmap
1768 root 20 0 1728 484 420 S 0 0.1 0:00.00 mingetty
1719 root 20 0 2384 452 176 S 0 0.1 0:00.00 dhclient
1626 daemon 20 0 2212 348 220 S 0 0.1 0:00.00 atd
2 root 20 0 0 0 0 S 0 0.0 0:00.00 kthreadd
3 root 20 0 0 0 0 S 0 0.0 0:00.00 ksoftirqd/0
4 root 20 0 0 0 0 S 0 0.0 0:00.00 kworker/0:0
5 root 0 -20 0 0 0 S 0 0.0 0:00.00 kworker/0:0H
6 root 20 0 0 0 0 S 0 0.0 0:00.02 kworker/u:0
7 root 0 -20 0 0 0 S 0 0.0 0:00.00 kworker/u:0H
8 root RT 0 0 0 0 S 0 0.0 0:00.00 migration/0
9 root 20 0 0 0 0 S 0 0.0 0:00.00 rcu_bh
10 root 20 0 0 0 0 S 0 0.0 0:00.01 rcu_sched
11 root 20 0 0 0 0 S 0 0.0 0:00.00 ksoftirqd/1
And htop output:
More than 400MB of hard drive space was used, which is quite standard:
df -h Filesystem Size Used Avail Use% Mounted on /dev/xvda 24G 439M 22G 2% / tmpfs 249M 0 249M 0% /lib/init/rw udev 10M 112K 9.9M 2% /dev tmpfs 249M 0 249M 0% /dev/shm
And the iNodes are obviously set according to the size of the hard drive:
df -i Filesystem Inodes IUsed IFree IUse% Mounted on /dev/xvda 2511040 22485 2488555 1% / tmpfs 63512 5 63507 1% /lib/init/rw udev 63460 2090 61370 4% /dev tmpfs 63512 1 63511 1% /dev/shm
When the full LNMP stack was loaded, more than 80MB of RAM was used, which is quite high:
free -m
total used free shared buffers cached
Mem: 496 475 20 0 33 358
-/+ buffers/cache: 82 413
Swap: 255 0 255
Top showing the full LNMP stack running:
top - 11:49:33 up 1:45, 1 user, load average: 9.15, 10.93, 6.41
Tasks: 91 total, 1 running, 90 sleeping, 0 stopped, 0 zombie
Cpu(s): 0.0%us, 0.0%sy, 0.0%ni,100.0%id, 0.0%wa, 0.0%hi, 0.0%si, 0.0%st
Mem: 508100k total, 486488k used, 21612k free, 34732k buffers
Swap: 262140k total, 56k used, 262084k free, 367616k cached
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
4755 www 20 0 15064 11m 672 S 0 2.2 0:00.02 nginx
4736 mysql 20 0 33992 4556 1980 S 0 0.9 0:00.00 mysqld
4746 root 20 0 24744 4548 1404 S 0 0.9 0:00.90 php-cgi
4747 www 20 0 24744 4068 924 S 0 0.8 0:00.00 php-cgi
4748 www 20 0 24744 4068 924 S 0 0.8 0:00.00 php-cgi
4750 www 20 0 24744 4068 924 S 0 0.8 0:00.00 php-cgi
4752 www 20 0 24744 4068 924 S 0 0.8 0:00.00 php-cgi
4753 www 20 0 24744 4068 924 S 0 0.8 0:00.00 php-cgi
1775 root 20 0 6088 2476 1152 S 0 0.5 0:00.25 bash
1772 root 20 0 8576 1612 960 S 0 0.3 0:02.09 sshd
1661 ntp 20 0 4712 1144 720 S 0 0.2 0:00.34 ntpd
6978 root 20 0 2500 1072 816 R 0 0.2 0:00.00 top
1597 root 20 0 28492 1012 664 S 0 0.2 0:00.00 rsyslogd
1750 root 20 0 5544 744 356 S 0 0.1 0:00.00 sshd
4754 root 20 0 4852 664 220 S 0 0.1 0:00.00 nginx
1087 root 16 -4 2308 556 228 S 0 0.1 0:00.06 udevd
1690 root 20 0 4396 552 388 S 0 0.1 0:00.00 cron
4634 root 20 0 1804 544 460 S 0 0.1 0:00.01 mysqld_safe
1 root 20 0 2068 532 456 S 0 0.1 0:01.45 init
1426 statd 20 0 1988 504 356 S 0 0.1 0:00.00 rpc.statd
6025 root 20 0 1732 492 424 S 0 0.1 0:00.00 mingetty
1158 root 18 -2 2304 480 156 S 0 0.1 0:00.00 udevd
1154 root 18 -2 2304 476 152 S 0 0.1 0:00.00 udevd
1414 daemon 20 0 1860 400 304 S 0 0.1 0:00.00 portmap
1719 root 20 0 2384 312 36 S 0 0.1 0:00.00 dhclient
1626 daemon 20 0 2212 308 180 S 0 0.1 0:00.00 atd
2 root 20 0 0 0 0 S 0 0.0 0:00.00 kthreadd
3 root 20 0 0 0 0 S 0 0.0 0:00.18 ksoftirqd/0
4 root 20 0 0 0 0 S 0 0.0 0:00.00 kworker/0:0
5 root 0 -20 0 0 0 S 0 0.0 0:00.00 kworker/0:0H
And again, htop output:
About 1.7GB of hard drive space was used after the installation, which is quite standard:
df -h Filesystem Size Used Avail Use% Mounted on /dev/xvda 24G 1.7G 21G 8% / tmpfs 249M 0 249M 0% /lib/init/rw udev 10M 112K 9.9M 2% /dev tmpfs 249M 0 249M 0% /dev/shm
As I was supposed to get the test done within 4 hours, I almost did not give this poor little machine a single second of break, and therefore the uptime might not be that great considering I have just run a Unixbench few minutes before:
uptime 11:51:41 up 1:47, 1 user, load average: 1.07, 7.09, 5.57
However, a little bit more of waiting and the results is a lot better:
uptime 12:10:22 up 2:06, 1 user, load average: 0.00, 0.17, 1.68
From vmstat, you can see at least there is no iowait, which is great:
vmstat procs -----------memory---------- ---swap-- -----io---- -system-- ----cpu---- r b swpd free buff cache si so bi bo in cs us sy id wa 0 0 56 21736 34756 367720 0 0 13 107 1223 655 14 19 66 0
Running vmstat again when I ran uptime command for the second time shows better results in CPU usage, but IOWait remains to be 0, which is good:
vmstat procs -----------memory---------- ---swap-- -----io---- -system-- ----cpu---- r b swpd free buff cache si so bi bo in cs us sy id wa 0 0 56 7236 20740 422268 0 0 12 163 1045 559 12 16 71 0
Meminfo shows there are four CPU cores assigned to the VPS, and none of them are throttled:
cat /proc/cpuinfo processor : 0 vendor_id : GenuineIntel cpu family : 6 model : 26 model name : Intel(R) Xeon(R) CPU L5520 @ 2.27GHz stepping : 5 microcode : 0x16 cpu MHz : 2266.842 cache size : 8192 KB physical id : 0 siblings : 4 core id : 0 cpu cores : 1 apicid : 0 initial apicid : 20 fdiv_bug : no hlt_bug : no f00f_bug : no coma_bug : no fpu : yes fpu_exception : yes cpuid level : 11 wp : yes flags : fpu de tsc msr pae cx8 cmov pat clflush mmx fxsr sse sse2 ss ht nx constant_tsc nonstop_tsc pni ssse3 sse4_1 sse4_2 popcnt hypervisor ida dtherm bogomips : 4533.68 clflush size : 64 cache_alignment : 64 address sizes : 40 bits physical, 48 bits virtual power management: processor : 1 vendor_id : GenuineIntel cpu family : 6 model : 26 model name : Intel(R) Xeon(R) CPU L5520 @ 2.27GHz stepping : 5 microcode : 0x16 cpu MHz : 2266.842 cache size : 8192 KB physical id : 0 siblings : 4 core id : 0 cpu cores : 1 apicid : 0 initial apicid : 20 fdiv_bug : no hlt_bug : no f00f_bug : no coma_bug : no fpu : yes fpu_exception : yes cpuid level : 11 wp : yes flags : fpu de tsc msr pae cx8 cmov pat clflush mmx fxsr sse sse2 ss ht nx constant_tsc nonstop_tsc pni ssse3 sse4_1 sse4_2 popcnt hypervisor ida dtherm bogomips : 4533.68 clflush size : 64 cache_alignment : 64 address sizes : 40 bits physical, 48 bits virtual power management: processor : 2 vendor_id : GenuineIntel cpu family : 6 model : 26 model name : Intel(R) Xeon(R) CPU L5520 @ 2.27GHz stepping : 5 microcode : 0x16 cpu MHz : 2266.842 cache size : 8192 KB physical id : 0 siblings : 4 core id : 0 cpu cores : 1 apicid : 0 initial apicid : 20 fdiv_bug : no hlt_bug : no f00f_bug : no coma_bug : no fpu : yes fpu_exception : yes cpuid level : 11 wp : yes flags : fpu de tsc msr pae cx8 cmov pat clflush mmx fxsr sse sse2 ss ht nx constant_tsc nonstop_tsc pni ssse3 sse4_1 sse4_2 popcnt hypervisor ida dtherm bogomips : 4533.68 clflush size : 64 cache_alignment : 64 address sizes : 40 bits physical, 48 bits virtual power management: processor : 3 vendor_id : GenuineIntel cpu family : 6 model : 26 model name : Intel(R) Xeon(R) CPU L5520 @ 2.27GHz stepping : 5 microcode : 0x16 cpu MHz : 2266.842 cache size : 8192 KB physical id : 0 siblings : 4 core id : 0 cpu cores : 1 apicid : 0 initial apicid : 20 fdiv_bug : no hlt_bug : no f00f_bug : no coma_bug : no fpu : yes fpu_exception : yes cpuid level : 11 wp : yes flags : fpu de tsc msr pae cx8 cmov pat clflush mmx fxsr sse sse2 ss ht nx constant_tsc nonstop_tsc pni ssse3 sse4_1 sse4_2 popcnt hypervisor ida dtherm bogomips : 4533.68 clflush size : 64 cache_alignment : 64 address sizes : 40 bits physical, 48 bits virtual power management:
It is interesting to note that Linode actually uses pretty old, energy-efficient CPUs like L5520 rather than the new E3 or E5 CPUs.
Meminfo output:
cat /proc/meminfo MemTotal: 508100 kB MemFree: 7244 kB Buffers: 20424 kB Cached: 422672 kB SwapCached: 24 kB Active: 43776 kB Inactive: 422428 kB Active(anon): 8884 kB Inactive(anon): 14312 kB Active(file): 34892 kB Inactive(file): 408116 kB Unevictable: 0 kB Mlocked: 0 kB HighTotal: 0 kB HighFree: 0 kB LowTotal: 508100 kB LowFree: 7244 kB SwapTotal: 262140 kB SwapFree: 262084 kB Dirty: 0 kB Writeback: 0 kB AnonPages: 23160 kB Mapped: 6604 kB Shmem: 68 kB Slab: 21156 kB SReclaimable: 14072 kB SUnreclaim: 7084 kB KernelStack: 768 kB PageTables: 804 kB NFS_Unstable: 0 kB Bounce: 0 kB WritebackTmp: 0 kB CommitLimit: 516188 kB Committed_AS: 105400 kB VmallocTotal: 329720 kB VmallocUsed: 2976 kB VmallocChunk: 326720 kB DirectMap4k: 532480 kB DirectMap2M: 0 kB
And time sync:
time sync real 0m0.003s user 0m0.000s sys 0m0.002s
The download speed is decent considering it is on a 100mbit port and the VPS is in Tokyo, which Cachefly probably does not have that much extensive network as what they have in North America:
wget cachefly.cachefly.net/100mb.test -O /dev/null --2013-03-09 11:56:08-- http://cachefly.cachefly.net/100mb.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: 104857600 (100M) [application/octet-stream] Saving to: a€?/dev/nulla€ 100%[=======================================================================================================================================>] 104,857,600 9.99M/s in 12s 2013-03-09 11:56:20 (8.60 MB/s) - a€?/dev/nulla€
Testing a little later again and the speed is slightly worse:
wget cachefly.cachefly.net/100mb.test -O /dev/null --2013-03-09 12:10:44-- http://cachefly.cachefly.net/100mb.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: 104857600 (100M) [application/octet-stream] Saving to: a€?/dev/nulla€ 100%[=======================================================================================================================================>] 104,857,600 8.38M/s in 13s 2013-03-09 12:10:57 (7.80 MB/s) - a€?/dev/nulla€
For the upload test, I did not expect much since we all know that in general, connectivity from Asia to the rest of the world is not too great, and indeed that turned out to be the case. When downloading to my Quickweb test VPS in Chicago, IL, I was only able to get close to 700KB/s:
wget 106.187.54.6/100mb.test -O /dev/null --2013-03-08 22:57:07-- http://106.187.54.6/100mb.test Connecting to 106.187.54.6:80... connected. HTTP request sent, awaiting response... 200 OK Length: 104857600 (100M) [application/octet-stream] Saving to: `/dev/null' 100%[======================================>] 104,857,600 696K/s in 2m 28s 2013-03-08 22:59:35 (691 KB/s) - `/dev/null' saved [104857600/104857600]
Downloading from the West Coast on my ChicagoVPS VPS in Los Angeles, IL, shows even worse result:
wget 106.187.54.6/100mb.test -O /dev/null --2013-03-09 08:02:11-- http://106.187.54.6/100mb.test Connecting to 106.187.54.6:80... connected. HTTP request sent, awaiting response... 200 OK Length: 104857600 (100M) [application/octet-stream] Saving to: `/dev/null' 100%[======================================>] 104,857,600 552K/s in 3m 23s 2013-03-09 08:05:34 (505 KB/s) - `/dev/null' saved [104857600/104857600]
And when I tried to download the test file from the test VPS from DireVPS in Maindenhead, UK, the speed is even worse at slightly more than 500KB/s:
wget 106.187.54.6/100mb.test -O /dev/null --2013-03-09 08:02:11-- http://106.187.54.6/100mb.test Connecting to 106.187.54.6:80... connected. HTTP request sent, awaiting response... 200 OK Length: 104857600 (100M) [application/octet-stream] Saving to: `/dev/null' 100%[======================================>] 104,857,600 552K/s in 3m 23s 2013-03-09 08:05:34 (505 KB/s) - `/dev/null' saved [104857600/104857600]
The only bright spot among all the speed tests is from my Prometeus test VPS located in Milan, Italy, and for some amazing reason, I was able to get to more than 5MB/s:
wget 106.187.54.6/100mb.test -O /dev/null --2013-03-08 23:06:12-- http://106.187.54.6/100mb.test Connecting to 106.187.54.6:80... connected. HTTP request sent, awaiting response... 200 OK Length: 104857600 (100M) [application/octet-stream] Saving to: `/dev/null' 100%[======================================>] 104,857,600 5.25M/s in 22s 2013-03-08 23:06:35 (4.51 MB/s) - `/dev/null' saved [104857600/104857600]
Since DireVPS and Prometeus are both located in continental Europe, I was actually quite curious why it was the case while I am writing these,and did a traceroute back to the Linode Tokyo VPS from both my VPS with DireVPS and Prometeus to see if there is any difference.
Turns out Prometeus actually has more hops to go through, but it is taking the Frankfurt -> New York -> San Jose -> Japan route:
traceroute to 106.187.54.6 (106.187.54.6), 30 hops max, 60 byte packets 1 pm17.prometeus.net (192.71.245.117) 0.036 ms 0.016 ms 0.014 ms 2 gw-cdlan-2.prometeus.cdlan.net (217.171.46.253) 0.409 ms 0.483 ms 0.558 ms 3 212.73.241.129 (212.73.241.129) 0.150 ms 0.164 ms 0.141 ms 4 ae-14-14.ebr1.Frankfurt1.Level3.net (4.69.142.194) 156.697 ms 156.762 ms 156.743 ms 5 ae-91-91.csw4.Frankfurt1.Level3.net (4.69.140.14) 159.771 ms ae-81-81.csw3.Frankfurt1.Level3.net (4.69.140.10) 156.510 ms ae-71-71.csw2.Frankfurt1.Level3.net (4.69.140.6) 156.135 ms 6 ae-92-92.ebr2.Frankfurt1.Level3.net (4.69.140.29) 163.318 ms ae-82-82.ebr2.Frankfurt1.Level3.net (4.69.140.25) 164.247 ms 164.170 ms 7 * * * 8 * * * 9 ae-71-71.csw2.NewYork1.Level3.net (4.69.134.70) 173.058 ms ae-61-61.csw1.NewYork1.Level3.net (4.69.134.66) 160.378 ms 160.272 ms 10 ae-72-72.ebr2.NewYork1.Level3.net (4.69.148.37) 164.656 ms 164.185 ms ae-62-62.ebr2.NewYork1.Level3.net (4.69.148.33) 158.105 ms 11 4.69.135.185 (4.69.135.185) 156.118 ms 156.296 ms 156.179 ms 12 ae-81-81.csw3.SanJose1.Level3.net (4.69.153.10) 155.730 ms ae-91-91.csw4.SanJose1.Level3.net (4.69.153.14) 158.815 ms ae-71-71.csw2.SanJose1.Level3.net (4.69.153.6) 156.205 ms 13 ae-3-80.edge1.SanJose2.Level3.net (4.69.152.135) 162.596 ms ae-1-60.edge1.SanJose2.Level3.net (4.69.152.7) 160.925 ms ae-3-80.edge1.SanJose2.Level3.net (4.69.152.135) 164.665 ms 14 KDDI-AMERIC.edge1.SanJose2.Level3.net (4.28.172.22) 165.296 ms 165.312 ms 164.558 ms 15 pajbb001.int-gw.kddi.ne.jp (111.87.3.49) 166.823 ms pajbb002.int-gw.kddi.ne.jp (111.87.3.53) 161.624 ms 160.963 ms 16 otejbb206.int-gw.kddi.ne.jp (203.181.100.153) 289.325 ms otejbb205.int-gw.kddi.ne.jp (203.181.100.137) 273.034 ms 273.359 ms 17 cm-fcu203.kddnet.ad.jp (124.215.194.180) 278.613 ms cm-fcu203.kddnet.ad.jp (124.215.194.164) 272.763 ms cm-fcu203.kddnet.ad.jp (124.215.194.180) 276.008 ms 18 124.215.199.122 (124.215.199.122) 278.101 ms 268.418 ms 277.754 ms
And with DireVPS, it was taking the London -> LA -> Japan route:
traceroute 106.187.54.6 traceroute to 106.187.54.6 (106.187.54.6), 30 hops max, 60 byte packets 1 12.108.255.149.in-addr.arpa (149.255.108.12) 0.036 ms 0.010 ms 0.007 ms 2 87.117.212.41 (87.117.212.41) 0.328 ms 0.318 ms 0.307 ms 3 610.core1.hex.as20860.net (62.233.127.181) 1.186 ms 1.211 ms 1.261 ms 4 TenGigabitEthernet7-1.ar4.LON3.gblx.net (64.211.1.189) 1.259 ms 1.298 ms 1.336 ms 5 ae0.scr3.LON3.gblx.net (67.16.142.101) 1.216 ms 1.207 ms 1.197 ms 6 xe2-3-2-10G.scr3.LAX1.gblx.net (67.16.165.65) 131.997 ms 132.008 ms 131.996 ms 7 e5-1-40G.ar6.LAX1.gblx.net (67.17.111.65) 139.965 ms 132.110 ms 132.107 ms 8 64.208.27.114 (64.208.27.114) 241.854 ms 241.961 ms 242.003 ms 9 lajbb001.int-gw.kddi.ne.jp (59.128.2.209) 255.244 ms 255.241 ms lajbb002.k 10 otejbb206.int-gw.kddi.ne.jp (203.181.100.21) 243.136 ms 243.139 ms 243.12 11 cm-fcu203.kddnet.ad.jp (124.215.194.164) 266.662 ms cm-fcu203.kddnet.ad.jp 12 124.215.199.122 (124.215.199.122) 268.107 ms 259.761 ms 259.760 ms
Since both GlobalCrossing (gblx.net) and Level3 (level3.net) are under the same brand and I would assume they have similar infrastructure (correct me if I am wrong please), I can only assume the route taken by Prometeus is a much (and apparently faster route) than GlobalCrossing.
The disk I/O is not the greatest but pretty decent:
dd if=/dev/zero of=test bs=64k count=16k conv=fdatasync 16384+0 records in 16384+0 records out 1073741824 bytes (1.1 GB) copied, 10.5379 s, 102 MB/s
Testing again showed similar results:
dd if=/dev/zero of=test bs=64k count=16k conv=fdatasync 16384+0 records in 16384+0 records out 1073741824 bytes (1.1 GB) copied, 10.0793 s, 107 MB/s
IOPing shows the disk is in general pretty stable, however it did flicker a little towards the end of my first ioping test on the VPS:
4096 bytes from . (ext3 /dev/root): request=1 time=0.1 ms 4096 bytes from . (ext3 /dev/root): request=2 time=0.2 ms 4096 bytes from . (ext3 /dev/root): request=3 time=0.2 ms 4096 bytes from . (ext3 /dev/root): request=4 time=0.2 ms 4096 bytes from . (ext3 /dev/root): request=5 time=0.2 ms 4096 bytes from . (ext3 /dev/root): request=6 time=0.2 ms 4096 bytes from . (ext3 /dev/root): request=7 time=0.2 ms 4096 bytes from . (ext3 /dev/root): request=8 time=0.2 ms 4096 bytes from . (ext3 /dev/root): request=9 time=0.2 ms 4096 bytes from . (ext3 /dev/root): request=10 time=65.5 ms --- . (ext3 /dev/root) ioping statistics --- 10 requests completed in 9068.8 ms, 150 iops, 0.6 mb/s min/avg/max/mdev = 0.1/6.7/65.5/19.6 ms
Testing again and the results are better:
dd if=/dev/zero of=test bs=64k count=16k conv=fdatasync 16384+0 records in 16384+0 records out 1073741824 bytes (1.1 GB) copied, 10.0793 s, 107 MB/s ioping -c 10 . 4096 bytes from . (ext3 /dev/root): request=1 time=0.1 ms 4096 bytes from . (ext3 /dev/root): request=2 time=0.2 ms 4096 bytes from . (ext3 /dev/root): request=3 time=0.2 ms 4096 bytes from . (ext3 /dev/root): request=4 time=0.1 ms 4096 bytes from . (ext3 /dev/root): request=5 time=0.2 ms 4096 bytes from . (ext3 /dev/root): request=6 time=0.2 ms 4096 bytes from . (ext3 /dev/root): request=7 time=0.2 ms 4096 bytes from . (ext3 /dev/root): request=8 time=0.2 ms 4096 bytes from . (ext3 /dev/root): request=9 time=0.2 ms 4096 bytes from . (ext3 /dev/root): request=10 time=0.2 ms --- . (ext3 /dev/root) ioping statistics --- 10 requests completed in 9003.9 ms, 5467 iops, 21.4 mb/s min/avg/max/mdev = 0.1/0.2/0.2/0.0 ms
Benchmark testing was what takes most of the time in general, so I was a little afraid that I did not get a chance to complete my usual benchmark testing (which is Unixbench twice, Geekbench twice). As such, I ran a Unixbench test when the VPS was first booted and after about an hour, this is what I see:
# # # # # # # ##### ###### # # #### # #
# # ## # # # # # # # ## # # # # #
# # # # # # ## ##### ##### # # # # ######
# # # # # # ## # # # # # # # # #
# # # ## # # # # # # # ## # # # #
#### # # # # # ##### ###### # # #### # #
Version 5.1.3 Based on the Byte Magazine Unix Benchmark
Multi-CPU version Version 5 revisions by Ian Smith,
Sunnyvale, CA, USA
January 13, 2011 johantheghost at yahoo period com
1 x Dhrystone 2 using register variables 1 2 3 4 5 6 7 8 9 10
1 x Double-Precision Whetstone 1 2 3 4 5 6 7 8 9 10
1 x Execl Throughput 1 2 3
1 x File Copy 1024 bufsize 2000 maxblocks 1 2 3
1 x File Copy 256 bufsize 500 maxblocks 1 2 3
1 x File Copy 4096 bufsize 8000 maxblocks 1 2 3
1 x Pipe Throughput 1 2 3 4 5 6 7 8 9 10
1 x Pipe-based Context Switching 1 2 3 4 5 6 7 8 9 10
1 x Process Creation 1 2 3
1 x System Call Overhead 1 2 3 4 5 6 7 8 9 10
1 x Shell Scripts (1 concurrent) 1 2 3
1 x Shell Scripts (8 concurrent) 1 2 3
4 x Dhrystone 2 using register variables 1 2 3 4 5 6 7 8 9 10
4 x Double-Precision Whetstone 1 2 3 4 5 6 7 8 9 10
4 x Execl Throughput 1 2 3
4 x File Copy 1024 bufsize 2000 maxblocks 1 2 3
4 x File Copy 256 bufsize 500 maxblocks 1 2 3
4 x File Copy 4096 bufsize 8000 maxblocks 1 2 3
4 x Pipe Throughput 1 2 3 4 5 6 7 8 9 10
4 x Pipe-based Context Switching 1 2 3 4 5 6 7 8 9 10
4 x Process Creation 1 2 3
4 x System Call Overhead 1 2 3 4 5 6 7 8 9 10
4 x Shell Scripts (1 concurrent) 1 2 3
4 x Shell Scripts (8 concurrent) 1 2 3
========================================================================
BYTE UNIX Benchmarks (Version 5.1.3)
System: li406-6: GNU/Linux
OS: GNU/Linux -- 3.7.10-linode49 -- #2 SMP Wed Feb 27 14:15:25 EST 2013
Machine: i686 (unknown)
Language: en_US.utf8 (charmap="UTF-8", collate="UTF-8")
CPU 0: Intel(R) Xeon(R) CPU L5520 @ 2.27GHz (4533.7 bogomips)
Hyper-Threading, MMX, Physical Address Ext
CPU 1: Intel(R) Xeon(R) CPU L5520 @ 2.27GHz (4533.7 bogomips)
Hyper-Threading, MMX, Physical Address Ext
CPU 2: Intel(R) Xeon(R) CPU L5520 @ 2.27GHz (4533.7 bogomips)
Hyper-Threading, MMX, Physical Address Ext
CPU 3: Intel(R) Xeon(R) CPU L5520 @ 2.27GHz (4533.7 bogomips)
Hyper-Threading, MMX, Physical Address Ext
10:52:39 up 48 min, 1 user, load average: 0.14, 1.07, 1.05; runlevel 2
------------------------------------------------------------------------
Benchmark Run: Sat Mar 09 2013 10:52:39 - 11:20:32
4 CPUs in system; running 1 parallel copy of tests
Dhrystone 2 using register variables 12107109.4 lps (10.0 s, 7 samples)
Double-Precision Whetstone 2282.4 MWIPS (10.3 s, 7 samples)
Execl Throughput 1673.2 lps (30.0 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks 415721.6 KBps (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks 110845.3 KBps (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks 1000748.6 KBps (30.0 s, 2 samples)
Pipe Throughput 619820.8 lps (10.0 s, 7 samples)
Pipe-based Context Switching 33203.0 lps (10.0 s, 7 samples)
Process Creation 2927.5 lps (30.0 s, 2 samples)
Shell Scripts (1 concurrent) 4154.8 lpm (60.0 s, 2 samples)
Shell Scripts (8 concurrent) 1274.3 lpm (60.0 s, 2 samples)
System Call Overhead 470786.1 lps (10.0 s, 7 samples)
System Benchmarks Index Values BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 12107109.4 1037.5
Double-Precision Whetstone 55.0 2282.4 415.0
Execl Throughput 43.0 1673.2 389.1
File Copy 1024 bufsize 2000 maxblocks 3960.0 415721.6 1049.8
File Copy 256 bufsize 500 maxblocks 1655.0 110845.3 669.8
File Copy 4096 bufsize 8000 maxblocks 5800.0 1000748.6 1725.4
Pipe Throughput 12440.0 619820.8 498.2
Pipe-based Context Switching 4000.0 33203.0 83.0
Process Creation 126.0 2927.5 232.3
Shell Scripts (1 concurrent) 42.4 4154.8 979.9
Shell Scripts (8 concurrent) 6.0 1274.3 2123.9
System Call Overhead 15000.0 470786.1 313.9
========
System Benchmarks Index Score 573.9
------------------------------------------------------------------------
Benchmark Run: Sat Mar 09 2013 11:20:32 - 11:48:28
4 CPUs in system; running 4 parallel copies of tests
Dhrystone 2 using register variables 41692937.3 lps (10.0 s, 7 samples)
Double-Precision Whetstone 8699.5 MWIPS (10.4 s, 7 samples)
Execl Throughput 5910.9 lps (30.0 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks 750775.3 KBps (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks 194696.9 KBps (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks 1997695.4 KBps (30.0 s, 2 samples)
Pipe Throughput 2291065.3 lps (10.0 s, 7 samples)
Pipe-based Context Switching 278929.9 lps (10.0 s, 7 samples)
Process Creation 10579.3 lps (30.0 s, 2 samples)
Shell Scripts (1 concurrent) 10814.3 lpm (60.0 s, 2 samples)
Shell Scripts (8 concurrent) 1393.5 lpm (60.1 s, 2 samples)
System Call Overhead 1760326.1 lps (10.0 s, 7 samples)
System Benchmarks Index Values BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 41692937.3 3572.7
Double-Precision Whetstone 55.0 8699.5 1581.7
Execl Throughput 43.0 5910.9 1374.6
File Copy 1024 bufsize 2000 maxblocks 3960.0 750775.3 1895.9
File Copy 256 bufsize 500 maxblocks 1655.0 194696.9 1176.4
File Copy 4096 bufsize 8000 maxblocks 5800.0 1997695.4 3444.3
Pipe Throughput 12440.0 2291065.3 1841.7
Pipe-based Context Switching 4000.0 278929.9 697.3
Process Creation 126.0 10579.3 839.6
Shell Scripts (1 concurrent) 42.4 10814.3 2550.5
Shell Scripts (8 concurrent) 6.0 1393.5 2322.5
System Call Overhead 15000.0 1760326.1 1173.6
========
System Benchmarks Index Score 1663.0
For a four-core CPU, I would say the score is definitely not on the impressive end of the scale.
I had some time before the trial VPS was turned off (in fact, it was the last thing I did before the VPS turned off) so I ran the Unixbench test again, and the score is even slightly worse than before:
# # # # # # # ##### ###### # # #### # #
# # ## # # # # # # # ## # # # # #
# # # # # # ## ##### ##### # # # # ######
# # # # # # ## # # # # # # # # #
# # # ## # # # # # # # ## # # # #
#### # # # # # ##### ###### # # #### # #
Version 5.1.3 Based on the Byte Magazine Unix Benchmark
Multi-CPU version Version 5 revisions by Ian Smith,
Sunnyvale, CA, USA
January 13, 2011 johantheghost at yahoo period com
1 x Dhrystone 2 using register variables 1 2 3 4 5 6 7 8 9 10
1 x Double-Precision Whetstone 1 2 3 4 5 6 7 8 9 10
1 x Execl Throughput 1 2 3
1 x File Copy 1024 bufsize 2000 maxblocks 1 2 3
1 x File Copy 256 bufsize 500 maxblocks 1 2 3
1 x File Copy 4096 bufsize 8000 maxblocks 1 2 3
1 x Pipe Throughput 1 2 3 4 5 6 7 8 9 10
1 x Pipe-based Context Switching 1 2 3 4 5 6 7 8 9 10
1 x Process Creation 1 2 3
1 x System Call Overhead 1 2 3 4 5 6 7 8 9 10
1 x Shell Scripts (1 concurrent) 1 2 3
1 x Shell Scripts (8 concurrent) 1 2 3
4 x Dhrystone 2 using register variables 1 2 3 4 5 6 7 8 9 10
4 x Double-Precision Whetstone 1 2 3 4 5 6 7 8 9 10
4 x Execl Throughput 1 2 3
4 x File Copy 1024 bufsize 2000 maxblocks 1 2 3
4 x File Copy 256 bufsize 500 maxblocks 1 2 3
4 x File Copy 4096 bufsize 8000 maxblocks 1 2 3
4 x Pipe Throughput 1 2 3 4 5 6 7 8 9 10
4 x Pipe-based Context Switching 1 2 3 4 5 6 7 8 9 10
4 x Process Creation 1 2 3
4 x System Call Overhead 1 2 3 4 5 6 7 8 9 10
4 x Shell Scripts (1 concurrent) 1 2 3
4 x Shell Scripts (8 concurrent) 1 2 3
========================================================================
BYTE UNIX Benchmarks (Version 5.1.3)
System: li406-6: GNU/Linux
OS: GNU/Linux -- 3.7.10-linode49 -- #2 SMP Wed Feb 27 14:15:25 EST 2013
Machine: i686 (unknown)
Language: en_US.utf8 (charmap="UTF-8", collate="UTF-8")
CPU 0: Intel(R) Xeon(R) CPU L5520 @ 2.27GHz (4533.7 bogomips)
Hyper-Threading, MMX, Physical Address Ext
CPU 1: Intel(R) Xeon(R) CPU L5520 @ 2.27GHz (4533.7 bogomips)
Hyper-Threading, MMX, Physical Address Ext
CPU 2: Intel(R) Xeon(R) CPU L5520 @ 2.27GHz (4533.7 bogomips)
Hyper-Threading, MMX, Physical Address Ext
CPU 3: Intel(R) Xeon(R) CPU L5520 @ 2.27GHz (4533.7 bogomips)
Hyper-Threading, MMX, Physical Address Ext
12:39:39 up 2:35, 1 user, load average: 0.00, 0.01, 0.28; runlevel 2
------------------------------------------------------------------------
Benchmark Run: Sat Mar 09 2013 12:39:39 - 13:07:31
4 CPUs in system; running 1 parallel copy of tests
Dhrystone 2 using register variables 12474842.4 lps (10.0 s, 7 samples)
Double-Precision Whetstone 2301.5 MWIPS (10.2 s, 7 samples)
Execl Throughput 1707.4 lps (30.0 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks 422687.3 KBps (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks 113581.4 KBps (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks 986131.9 KBps (30.0 s, 2 samples)
Pipe Throughput 623029.0 lps (10.0 s, 7 samples)
Pipe-based Context Switching 33192.5 lps (10.0 s, 7 samples)
Process Creation 2998.7 lps (30.0 s, 2 samples)
Shell Scripts (1 concurrent) 4194.7 lpm (60.0 s, 2 samples)
Shell Scripts (8 concurrent) 1265.2 lpm (60.0 s, 2 samples)
System Call Overhead 471201.8 lps (10.0 s, 7 samples)
System Benchmarks Index Values BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 12474842.4 1069.0
Double-Precision Whetstone 55.0 2301.5 418.5
Execl Throughput 43.0 1707.4 397.1
File Copy 1024 bufsize 2000 maxblocks 3960.0 422687.3 1067.4
File Copy 256 bufsize 500 maxblocks 1655.0 113581.4 686.3
File Copy 4096 bufsize 8000 maxblocks 5800.0 986131.9 1700.2
Pipe Throughput 12440.0 623029.0 500.8
Pipe-based Context Switching 4000.0 33192.5 83.0
Process Creation 126.0 2998.7 238.0
Shell Scripts (1 concurrent) 42.4 4194.7 989.3
Shell Scripts (8 concurrent) 6.0 1265.2 2108.7
System Call Overhead 15000.0 471201.8 314.1
========
System Benchmarks Index Score 579.5
------------------------------------------------------------------------
Benchmark Run: Sat Mar 09 2013 13:07:31 - 13:35:24
4 CPUs in system; running 4 parallel copies of tests
Dhrystone 2 using register variables 42405873.1 lps (10.0 s, 7 samples)
Double-Precision Whetstone 8725.8 MWIPS (10.2 s, 7 samples)
Execl Throughput 5945.3 lps (30.0 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks 639121.4 KBps (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks 195722.4 KBps (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks 1829198.7 KBps (30.0 s, 2 samples)
Pipe Throughput 2198171.9 lps (10.0 s, 7 samples)
Pipe-based Context Switching 263732.4 lps (10.0 s, 7 samples)
Process Creation 9621.8 lps (30.0 s, 2 samples)
Shell Scripts (1 concurrent) 10388.6 lpm (60.0 s, 2 samples)
Shell Scripts (8 concurrent) 1359.3 lpm (60.1 s, 2 samples)
System Call Overhead 1724663.1 lps (10.0 s, 7 samples)
System Benchmarks Index Values BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 42405873.1 3633.8
Double-Precision Whetstone 55.0 8725.8 1586.5
Execl Throughput 43.0 5945.3 1382.6
File Copy 1024 bufsize 2000 maxblocks 3960.0 639121.4 1613.9
File Copy 256 bufsize 500 maxblocks 1655.0 195722.4 1182.6
File Copy 4096 bufsize 8000 maxblocks 5800.0 1829198.7 3153.8
Pipe Throughput 12440.0 2198171.9 1767.0
Pipe-based Context Switching 4000.0 263732.4 659.3
Process Creation 126.0 9621.8 763.6
Shell Scripts (1 concurrent) 42.4 10388.6 2450.2
Shell Scripts (8 concurrent) 6.0 1359.3 2265.5
System Call Overhead 15000.0 1724663.1 1149.8
========
System Benchmarks Index Score 1595.7
Geekbench score, on the other hand, showed a somewhat more reasonable score:
System Information
Platform: Linux x86 (32-bit)
Compiler: GCC 4.1.2 20070925 (Red Hat 4.1.2-33)
Operating System: Linux 3.7.10-linode49 i686
Model: Linux PC (Intel(R) Xeon(R) CPU L5520 @ 2.27GHz)
Motherboard: Unknown Motherboard
Processor: Intel(R) Xeon(R) CPU L5520 @ 2.27GHz
Processor ID: GenuineIntel Family 6 Model 26 Stepping 5
Logical Processors: 4
Physical Processors: 1
Processor Frequency: 2.27 GHz
L1 Instruction Cache: 0.00 B
L1 Data Cache: 0.00 B
L2 Cache: 256 KB
L3 Cache: 0.00 B
Bus Frequency: 0.00 Hz
Memory: 496 MB
Memory Type: N/A
SIMD: 1
BIOS: N/A
Processor Model: Intel(R) Xeon(R) CPU L5520 @ 2.27GHz
Processor Cores: 4
Integer
Blowfish
single-threaded scalar 1555 ||||||
multi-threaded scalar 5996 |||||||||||||||||||||||
Text Compress
single-threaded scalar 1728 ||||||
multi-threaded scalar 5475 |||||||||||||||||||||
Text Decompress
single-threaded scalar 1793 |||||||
multi-threaded scalar 6097 ||||||||||||||||||||||||
Image Compress
single-threaded scalar 1475 |||||
multi-threaded scalar 4996 |||||||||||||||||||
Image Decompress
single-threaded scalar 1339 |||||
multi-threaded scalar 3991 |||||||||||||||
Lua
single-threaded scalar 2706 ||||||||||
multi-threaded scalar 7592 ||||||||||||||||||||||||||||||
Floating Point
Mandelbrot
single-threaded scalar 1879 |||||||
multi-threaded scalar 7346 |||||||||||||||||||||||||||||
Dot Product
single-threaded scalar 3028 ||||||||||||
multi-threaded scalar 9653 ||||||||||||||||||||||||||||||||||||||
single-threaded vector 3492 |||||||||||||
multi-threaded vector 11019 ||||||||||||||||||||||||||||||||||||||||||||
LU Decomposition
single-threaded scalar 2099 ||||||||
multi-threaded scalar 5211 ||||||||||||||||||||
Primality Test
single-threaded scalar 3026 ||||||||||||
multi-threaded scalar 8611 ||||||||||||||||||||||||||||||||||
Sharpen Image
single-threaded scalar 5065 ||||||||||||||||||||
multi-threaded scalar 13969 |||||||||||||||||||||||||||||||||||||||||||||||||||||||
Blur Image
single-threaded scalar 4605 ||||||||||||||||||
multi-threaded scalar 14928 |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Memory
Read Sequential
single-threaded scalar 4274 |||||||||||||||||
Write Sequential
single-threaded scalar 6100 ||||||||||||||||||||||||
Stdlib Allocate
single-threaded scalar 3289 |||||||||||||
Stdlib Write
single-threaded scalar 3337 |||||||||||||
Stdlib Copy
single-threaded scalar 6627 ||||||||||||||||||||||||||
Stream
Stream Copy
single-threaded scalar 3032 ||||||||||||
single-threaded vector 3791 |||||||||||||||
Stream Scale
single-threaded scalar 3230 ||||||||||||
single-threaded vector 3599 ||||||||||||||
Stream Add
single-threaded scalar 2770 |||||||||||
single-threaded vector 3497 |||||||||||||
Stream Triad
single-threaded scalar 2953 |||||||||||
single-threaded vector 2532 ||||||||||
Integer Score: 3728 ||||||||||||||
Floating Point Score: 6709 ||||||||||||||||||||||||||
Memory Score: 4725 ||||||||||||||||||
Stream Score: 3175 ||||||||||||
Overall Geekbench Score: 4915 |||||||||||||||||||
Testing again showed slightly better results:
System Information
Platform: Linux x86 (32-bit)
Compiler: GCC 4.1.2 20070925 (Red Hat 4.1.2-33)
Operating System: Linux 3.7.10-linode49 i686
Model: Linux PC (Intel(R) Xeon(R) CPU L5520 @ 2.27GHz)
Motherboard: Unknown Motherboard
Processor: Intel(R) Xeon(R) CPU L5520 @ 2.27GHz
Processor ID: GenuineIntel Family 6 Model 26 Stepping 5
Logical Processors: 4
Physical Processors: 1
Processor Frequency: 2.27 GHz
L1 Instruction Cache: 0.00 B
L1 Data Cache: 0.00 B
L2 Cache: 256 KB
L3 Cache: 0.00 B
Bus Frequency: 0.00 Hz
Memory: 496 MB
Memory Type: N/A
SIMD: 1
BIOS: N/A
Processor Model: Intel(R) Xeon(R) CPU L5520 @ 2.27GHz
Processor Cores: 4
Integer
Blowfish
single-threaded scalar 1564 ||||||
multi-threaded scalar 6523 ||||||||||||||||||||||||||
Text Compress
single-threaded scalar 1762 |||||||
multi-threaded scalar 6346 |||||||||||||||||||||||||
Text Decompress
single-threaded scalar 1790 |||||||
multi-threaded scalar 6911 |||||||||||||||||||||||||||
Image Compress
single-threaded scalar 1485 |||||
multi-threaded scalar 5516 ||||||||||||||||||||||
Image Decompress
single-threaded scalar 1317 |||||
multi-threaded scalar 4586 ||||||||||||||||||
Lua
single-threaded scalar 2713 ||||||||||
multi-threaded scalar 7688 ||||||||||||||||||||||||||||||
Floating Point
Mandelbrot
single-threaded scalar 1877 |||||||
multi-threaded scalar 7439 |||||||||||||||||||||||||||||
Dot Product
single-threaded scalar 3041 ||||||||||||
multi-threaded scalar 10933 |||||||||||||||||||||||||||||||||||||||||||
single-threaded vector 3629 ||||||||||||||
multi-threaded vector 15689 ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
LU Decomposition
single-threaded scalar 2129 ||||||||
multi-threaded scalar 6728 ||||||||||||||||||||||||||
Primality Test
single-threaded scalar 3066 ||||||||||||
multi-threaded scalar 8981 |||||||||||||||||||||||||||||||||||
Sharpen Image
single-threaded scalar 4999 |||||||||||||||||||
multi-threaded scalar 16637 ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Blur Image
single-threaded scalar 4661 ||||||||||||||||||
multi-threaded scalar 14020 ||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Memory
Read Sequential
single-threaded scalar 4226 ||||||||||||||||
Write Sequential
single-threaded scalar 6051 ||||||||||||||||||||||||
Stdlib Allocate
single-threaded scalar 3278 |||||||||||||
Stdlib Write
single-threaded scalar 3298 |||||||||||||
Stdlib Copy
single-threaded scalar 6532 ||||||||||||||||||||||||||
Stream
Stream Copy
single-threaded scalar 3286 |||||||||||||
single-threaded vector 4027 ||||||||||||||||
Stream Scale
single-threaded scalar 3192 ||||||||||||
single-threaded vector 3221 ||||||||||||
Stream Add
single-threaded scalar 2756 |||||||||||
single-threaded vector 3430 |||||||||||||
Stream Triad
single-threaded scalar 2878 |||||||||||
single-threaded vector 2557 ||||||||||
Integer Score: 4016 ||||||||||||||||
Floating Point Score: 7416 |||||||||||||||||||||||||||||
Memory Score: 4677 ||||||||||||||||||
Stream Score: 3168 ||||||||||||
Overall Geekbench Score: 5253 |||||||||||||||||||||
Customer Service and Support
One of the most remarkable feature that really impressed me was how fast Linode was when it comes to responding to support tickets. I raised a ticket about how to enter rDNS and it was responded in 28 seconds! Even though they do not have a fancy live chat feature on their support, their response to ticket is definitely close to instant:
On top of a great ticket response time, Linode also offers extensive resources for you to figure out the problem/do the work yourself, which includes the FAQ, Library, Forum, and IRC. In fact, Linode library is probably one of the most extensive library showing how to set up various software stack yourself, and the best part is it is actually open to public, meaning that you do not have to be a Linode member to enjoy this knowledgebase.
Finally, Linode has a very good notification system that keeps your VPS monitored. In the short four hours span, I have 12 emails from Linode that includes one welcome email, 4 ticket-related email (open, two replies and close), one disk I/O usage warning and two CPU warning, as well as four more emails whenever there are major events happened to the VPS (initial configuration, snapshot, shutdown and finally VPS termination). Basically for Linode, you can pretty much just let the server run without worrying too much.
Note that Linode also has its own IPhone app which you can use to monitor the VPS on your Apple devices, although it would be great if they could monitor the VPS on other mobile platform as well.
Conclusion
I found it is really difficult to give a one line comment about Linode. On one hand, we all know the prices it charges is not cheap and considering the old hardware and not-so-impressive network on the VPS, it probably does not worth the price. However, the support and the control panel, which are two major selling points of Linode, are indeed nothing less than absolutely awesome. Therefore, if you would like to have a VPS that you don’t have to worry about too much (until of course, your site gets DDOSed) once it is up and running, and get support whenever you need, then Linode would probably the best fit for you, otherwise, I have no doubt that you would be able to pay a little less and get a lot more.
Finally, what do you think of extreme reviews like this? Granted reviews like this are really quick and I might just be lucky to be on a good node and when the node is doing great, but is it something that would at least help? If so I’ll do a few more when I have time. Feel free to post your comments below. Thanks!
Nodebalencers is a load balancing service not a VPS. I know you can achieve the same with a VPS but nodebalencers got less overhead and easy to use
@Ruchira: thanks for the clarification, so essentially you don’t get root access with the node balancers?
yep no root access. its just the control panel which you can enter the ips,ports of your servers and it just do the work for you 🙂
Ah, I see, so basically it will know how to mirror the contents automatically, that’s pretty good.
I did the 4 hour trial as well and received a VPS on a newer node. I can’t recall if it was E3 or E5 but it was one of the two.
Disk IO was better.
Though to be fair I think people who buy Linodes do so because of the network and service. Things like good staff, The ability to load balance. high uptime etc.
Its surprising that they have anythig other than L5520. L5520 is the only processor they have used for past 2+ years.
@Luma: What can I say? you are just one of their favorite boys 🙂 I could not capture the uptime since I only had it for 4 hours, but indeed the support really impressed me.
Great review.
I think when you really think about this review as a whole the reason Linode make it work the money is not that they use beading edge CPU’s, the greatest disk arrays, or even run the worlds fastest network it is that the quality of the ‘service’ is through the roof.
They have unique services, top level support staff, and a great reputation. I know 5520’s are not top of the line but let us be honest if you are using a VPS you don’t need dedicated hardware to begin with and you are unlikely to even use even 20% of a 5520 at any given time and if you really do need that you should not be using a VPS.
I think Linode just have things right.. they know what actually works and they know how to do things right and they know how to make sure you have everything you need from the 1st second with them.
As a VPS host this really makes me think hard about things, is the highest unixbench score important or would my efforts be better spent in developing service features and quality while providing a fair amount of performance.
My own reviews of Inception Hosting that you have done in the past blow the raw numbers out of the water, but can you really feel the difference between a 2000 and 1600 unixBench score as an end user? probably not, and can the avg user hitting your site tell the difference… almost certainly not.
Very interesting stuff, glad you did this.
Superb site you have here but I was curious if youu knew of any message boards that cover the same topics discussed in this article?
I’d really like to be a part of group where I can get responses from other experienced people that share the same interest.
If you have any recommendations, please leet me know.
Thanks!
@Edgar: Low end talk (lowendtalk.com) might be a good point to start if you are looking for more discussions on the low end VPS. 96MB also has its own forum (96forum.com) but it has been a while since anyone posted anything on it.