หนึ่งในโครงการที่ผมได้ทำเมื่อเร็ว ๆ นี้คือการเปรียบเทียบของ บริษัท ที่'ความนิยม'10 ให้บริการ VPS งบประมาณ ผมคิดว่าที่ผมได้ทำมาแล้วว่ามันจะเป็นสิ่งที่ดีที่จะให้กลับไปยังชุมชนและแบ่งปันประสบการณ์ของฉัน
ทางเลือกของบรรดา 10 บริษัท ขึ้นอยู่กับการกำหนดราคาของพวกเขาและการแสดงความคิดเห็นใน WebHostingTalk Lowendbox ประวัติและอื่น ๆ เนื่องจากมีเวลา จำกัด และงบประมาณของผมโชคไม่ดีเท่านั้นที่สามารถทดสอบ 10 (8 จริงเพราะของบางอย่างถูกหมดสต็อก), ดังนั้นหาก บริษัท ของคุณไม่ได้ที่นี่และคุณคิดว่าคุณมีคุณสมบัติตามเกณฑ์ซึ่งจะถูกระบุไว้ด้านล่างกรุณาติดต่อเรา
รายงานนี้จะไม่ได้มีรูปแบบมืออาชีพอย่างเคร่งครัดและผมจะพยายามให้สั้นและกระชับ ฉันจะไม่เกินวิเคราะห์มันไม่ได้สร้างสนามการขายใด ๆ เพียงแค่พยายามที่จะแบ่งปันประสบการณ์ของฉันซึ่งจะสามารถช่วยให้คุณ
One of the projects that I have recently done was a comparison of 10 ‘most popular’ companies providing budget VPS services. I thought that as I have already done that, it would be good to give something back to the community and share my experiences.
Choice of those 10 companies was based on their pricing, reviews on WebHostingTalk and Lowendbox, history and etc. Due to limited amount of time and budget I was unfortunately only able to test those 10 (actually 8 because of some being out of stock), so if your company is not here and you think that you meet the criteria, which will be listed below, please contact me.
This report will not have a strictly professional form and I will try to keep it short and concise. I will not over analyse it, not create any sales pitch just try to share my experiences, which potentially can help you.
About me:
To put the tests into a perspective I think it would be good to write at least a few sentences about me. I have been working few years as a network administrator (Linux) and then as a software/product developer (PHP/Python/C#.NET). Currently I am a freelancer and I am starting my own business.
Criteria for choosing companies:
512MB of guaranteed ram
Price below 8$ (some exceptions were allowed)
Minimum 10GB of disk space
At least reasonable reviews at Webhostingtalk / Lowendbox
Being in business for at least a year, preferably more
Possibility of upgrading the package without needing to pay a lot
Companies that were tested:
Quickweb
Clubuptime
Yard VPS
BurstNet
BuyVM
YDGH
Thrust VPS
Chicago VPS
A 512mb guaranteed memory package was bought from each. All of them except for Thrust VPS and Chicago VPS were based on OpenVZ platform. Chicago VPS did not provide OpenVZ and price difference on Thrust VPS was so low that I decided to put those two into tests as Xen. It is also worth mentioning that due to diversity, I took VPS from BurstNet and ThrustVPS in the UK.
Below, there is a table shortly comparing what sales pages have to say. Some prices already include the discount that was available during the time that I was buying them.
Code:
HOST RAM Guaranteed Ram Burst/Swap HDD BW IP Cpu Cores OS Price
Quickweb 512MB 768MB 25GB 250GB 1 1 Linux $6.95
Clubuptime 512MB 1024MB 10GB 100GB 1 1 Linux $3.98
Yard VPS 512MB 1024MB 30GB 1TB 1 2 Linux $5.95
BurstNet 512MB - 20GB 1TB 2 1 Linux $5.95
BuyVM 512MB 1024MB 50GB 2TB 1 2 Linux $5.95
YDGH 512MB - 20GB 1TB 1
[email protected] Linux $4.5
Thrust VPS 512MB - 15GB 1TB 1 4?! Linux $7.95
ChicagoVPS 512MB +256MB 20GB 1TB 2 1 Linux $6
As you can see from the table most of them look very comparable … but is it so in reality? Let’s try to answer that question.
About tests:
Probably I should write here what tests were performed, why and etc … Instead of that I will just jump straight into them =).
1) Network Tests
For the time being I have performed only limited network testing. This was based on download speed from the following locations:
http://cachefly.cachefly.net/100mb.test 
http://ftp://ftp.ovh.net/test.bin http://85.17.128.2/100mb.bin (leaseweb)
http://speedtest.dal01.softlayer.com ...ds/test100.zip - dallas
http://speedtest.wdc01.softlayer.com ...ds/test100.zip - washington
http://www.singlehop.com/speedtest/100megabytefile I tried to base it on rather popular and good networks (both USA and Europe). The below table summarizes the results (speed in MB/s).
Code:
HOST Cachefly OVH Leaseweb Softlayer Dallas Softlayer Washington Singlehop Port speed
Quickweb 69 8.2 7.03 28.6 12 10.5 1gbps
Clubuptime 56 9.7 9.4 108 24 11 1gbps
Yard VPS 7 1.76 1.1 1.9 1.33 3.3 100mbps
BurstNet 22 40.7 60 7 9.9 5 1gbps
BuyVM 25 3.2 0.35 21.5 10 7.5 1gbps
YDGH 10 4.9 5 10.1 9.8 9.7 100mbps
Thrust VPS 7.5 4.1 3.8 0.5 0.7 5.7 100mbps
ChicagoVPS 11.2 5.1 4.21 11.2 10.5 11.2 100mbps
The companies were put here in a no particular order. I will only name 3 companies that had fastest speeds, those were: BurstNet, Clubuptime and Quickweb.
2) I/O Test:
These tests were measured by sequential write repeated many times at different days and hours.
dd if=/dev/zero of=test bs=64k count=16k conv=fdatasync
And by installing different packages (for bigger packages, the download speeds were subtracted).
time apt-get -y install openoffice.org
time apt-get -y install mc
time apt-get -y install chromium-browser
time apt-get -y install openarena (because its large ! =) )
Again results were compiled into a table:
Code:
HOST dd (MB/s) average dd (MB/s) Installing OO Installing MC Installing chrome Installing Open Arena
Quickweb 25-140 85 1.09m 7.1s 8.4s 12s
Clubuptime 95-170 115 0.49m 5.1s 8.8s 9s
Yard VPS 5-50 15 59 MINUTES!!! - - -
BurstNet 10-22 17 4.13m 28.8s 45s 80s
BuyVM 19-135 75 1.54m 7.1s 19.7s 15s
YDGH 15 or 60-105 15 or 70 1.41/3.50m 9.9/14s 15/32s 20/57s
Thrust VPS 10-65 30 3.39m 6.3s 12.2s 42s
ChicagoVPS 70-80 75 1.34m 6.34s 14.1s 18.2s
First the list of top 3 performers:
Clubuptime
Quickweb
BuyVM / ChicagoVPs / YDGH
Few notes are needed here.
Yard VPS speed is TERRIBLE. I cannot stress that enough. Tried to contact them through ticket (critical level sitting for already FEW DAYS), through WHT and MSN. Completely no reaction. This is why after the first test I did not continue with others. This could happen to ANY other customer so choose them at your own risk, seriously.
YGDH has two results because it offers two servers. One for general use and one for users that need high I/O (although its CPU score will be lower).
3) Overall performance / CPU -> WHT 5.1.2 Benchmark
Overall performance was tested with WHT 5.1.2 Benchmark. I will also include what hardware the node is based on.
Code:
HOST WHT Score How many cores? Hardware -> CPU
Quickweb 817 1 X3440 @ 2.53
Clubuptime 1077 1 X3470 @ 2.93
Yard VPS 850@1, 1250@2 2 E5620 @ 1.8
BurstNet 764 1 ?! ss4.1 6mb cache 2.3ghz
BuyVM 740@1, 1141@2 2 e5520 @ 2.27
YDGH 404 I/O,601 general 1 Q9300 @1.1 / q9550 1.2ghz
Thrust VPS 790@1, 2150@4 4 E5620 @ 2.4
ChicagoVPS 460 1 E5405 @ 22
Again few notes are probably needed. Despite Yard VPS in theory scoring quite good in WHT benchmark, working with it is really not possible. Thrust VPS also seems to have GREAT results BUT when you look at different tasks as I/O, installing apps and etc you will see that it no longer looks so good .
4) Are they overselling?
You can never be sure! =). Lowendbox admin posted some time ago a good script that checks number of pages in OpenVZ and calculates how much memory is swapped. On XEN a driver called balloon is used to dynamically change the memory allocation of a VPS.
Code:
Host Result
Quickweb Swap: 0.00M (0.0% of Committed)
Clubuptime Swap: 0.00M (0.0% of Committed)
Yard VPS Swap: 0.00M (0.0% of Committed)
BurstNet Swap: 0.06M (0.1% of Committed)
BuyVM Swap: 0.07M (0.1% of Committed)
YDGH Swap: 0.00M (0.0% of Committed)
Thrust VPS XEN - no balloon seen
ChicagoVPS XEN - yes, there is balloon
I am not an expert in this field so please go to lowendbox and read the posts about overselling if you want some more information =). Basically some of the memory at BurstNet and BuyVM goes to swap.
The balloon gives ChicagoVPS a possibility to alter the 'dedicated' memory but at the moment they are not doing it:
#cat /proc/xen/balloon
Current allocation: 524288 kB
Requested target: 524288 kB
Minimum target: 173056 kB
Maximum target: 532480 kB
Low-mem balloon: 8192 kB
High-mem balloon: 0 kB
Driver pages: 232 kB
5) Pingdom – stability
I will not post the pingdom statistics yet – I prefer to do that at least at the end of the billing cycle. Although it is good to reserve the space and tell you that it is running! =).
6) Support times and handling
I should probably start here from the worst to the best.
Yard VPS – support? What support ?! Critical ticket sitting there for 4 days with completely no resolution. Alternative methods of contact – PMed Jimmy as well as trying MSN – no response despite him being available many times.
Summary : EXTREMELY BAD
Burst NET – I do not know if I should lough here or cry. Such a big company, the surely have a great support don’t they? Oh my, how wrong could I be!
I will try to shortly present the whole ticket history with them. Hours were always counted from the previous ticket.
Quote:
Me: Could you please check the VPS performance? <insert tests like dd, compare to others)
BurstNet (+6 hours): I am not seeing any issues. Do you still have them?
Me: Yes ! (tests)
BurstNet (+3 hours): Still not seeing any
Me: Strange because I can still see them < insert tests>
BurstNet (+2.5 hours): Your I/O problems may be caused by … resolv.conf configuration! Please give us SSH details
Me: <details> <tests>
BurstNet (+11 hours): Do you still have them?
// MY GOD ARE YOU SERIOUS? No my hobby is submitting tickets, I complain for fun.
Me: Yes. <tests> <installation of openoffice (300mb) at 4.30 minutes, dd 13.6mb/s>
BurstNet (+10 minutes), Sorry, I have to quote that one:
“I just tested this with another command and it completed in 18.35 seconds:
root@30000829:~# time apt-get -y install lynx
real 0m18.355s
user 0m2.868s
sys 0m2.826s
root@30000829:~# “
// This is getting even better! A 300KB package (lynx) has shorter installation time than a 300MB open office – what a surprise! Was not expecting that one! =))
Me: Really? I have tested them at few other VPS, e.g.:
real 0m3.007s
user 0m0.531s
sys 0m0.181s
BurstNet (+15 hours):
“root@30000829:/# time apt-get -y install openoffice.org
real 0m7.181s
user 0m1.449s
sys 0m2.115s
This is what I am seeing now for this vps. “
// You must be kidding me right? You tried installing an already installed package and hey, it took only 7 seconds! That’s a speedy machine!
This continues till now .
Time taken: 3days, Results: …. None?
Comment: At least they have good network, right? … NO! VERY BAD
Thrust VPS
This one is interesting as well! Although better than the previous ones.
Ticket created: on 5th Dec, 7:31PM
Quote:
Me: Could you please check performance <tests>
Thrust VPS (+15 Minutes): I am checking and will update you on this
Thrust VPS (+2 hours): Could you provide root details
Me: <details> <tests>
Thrust VPS (+4 minutes): I will update you on this.
Thrust VPS (+3 hours): Still updating you on this.
Thrust VPS (+3 hours): Im getting now 50mb/s. It’s all good now.
// Realllly? Peak times went into off-peak times and voila, magically it repaired itself (if we consider 40-50mb/s a good I/O). Which honestly is not bad.
Me: (closer to peak) – 15-18mb/s, very bad installation times/slowness while using
….
….
….
Many like these which is pointless to repeat here
Finally, after 3 days they think of moving me to another node.
I can’t complain here about them SO MUCH. They responded to the ticket in a reasonable time frame (sometimes even very fast). 50MB/s is not a bad I/O – I was complaining when it was in the regions of 15MB/s or even below 10MB/s for longer periods.
I know it is shared environment and it heavily differs what other people on the node are doing … but some companies manage to control it.
I should now probably comment about YDGH, Quickweb, BuyVM, Clubuptime, ChicagoVPS, but I will not. Why? I submitted few tickets to all of them and guess what – if there was a problem, they admitted it! They tried to explain why was it caused … they actually paid attention to a customer. It is hard to distinguish now that one had 30 minutes response time, other 1 hour and etc – it is sometimes based purely on luck. But I can assure those were definitely reasonable times and every issue (even one made up for the purposes of a ticket!) was explained/resolved quickly.
7) Summary
I am not with those hosts for long period so I cannot say too much. I can present you my results, my experiences and how do I feel about them. It is your decision which to choose or if you should judge them based on this report. I will try to update it when anything happens, when reasonable pingdom results are available and etc.
Which hosts I would strongly recommend? Clubuptime and Quickweb – they had constant great performance, treat their customers really good despite the low price tag and are in the market for quite many years.
What about BuyVM, YDGH, Chicago VPS? I would still recommend them, although due to a little worse performance or some issues (that were in the end resolved) I did not have so strong feelings about this as in the previous situation. But you should not be disappointed with them. There are also great guys working there.
What about Yard VPS, BurstNET and Thrust VPS? First thought is that putting Thrust in the same ‘bag’ as Yard and Burst is not fair for them . They at least tried to do something and are still trying to solve the ‘issue’. I will not say anything about Yard – I think that everything was already said ... I can just say that I, personally, will definitely AVOID them. What about BurstNet? They have good network that’s for sure … even a very good one … but the question is if that network is worth having poor (at least from my experience) support and not the best performance? I do not think so.
At the end I would just like to say – I wrote this short article because I thought it would be beneficial for users that want a budget VPS server. I bet some users could have completely different experiences than me. I can unfortunately only write based on my ‘journey’ with those companies.
I believe that people have a free choice, so read some more about those companies, speak with people that have VPS already there and you can of course use the material here to help you in the decision. It is yours after all .
<<snipped>>
I hope that at least parts of this article were useful to you or made you smile . On most web forums that I read I was always passive, so I figured out it would be good to at least give something back.
If you have any questions, write them here, I will try to answer them.
7) สรุปผมไม่ได้อยู่กับครอบครัวเหล่านั้นเป็นระยะเวลานานดังนั้นผมจึงไม่สามารถพูดมากเกินไป ฉันสามารถนำเสนอคุณผล, ประสบการณ์ของฉันฉันและฉันจะรู้สึกเกี่ยวกับพวกเขา มันคือการตัดสินใจของคุณที่จะเลือกหรือหากคุณควรจะตัดสินพวกเขาขึ้นอยู่กับรายงานนี้ ผมจะพยายามอัปเดตเมื่ออะไรจะเกิดขึ้นเมื่อผล Pingdom ที่เหมาะสมจะมีครอบครัวและอื่น ๆ ซึ่งผมขออยากจะแนะนำ? Clubuptime และ Quickweb -- พวกเขามีประสิทธิภาพที่ดีคงปฏิบัติต่อลูกค้าของพวกเขาดีจริงๆแม้จะมีป้ายราคาต่ำและอยู่ในตลาดเป็นเวลาหลายปีทีเดียว
สิ่งที่เกี่ยวกับ BuyVM, YDGH, Chicago VPS? ฉันยังจะขอแนะนำให้พวกเขา แต่เนื่องจากประสิทธิภาพการทำงานแย่ลงเพียงเล็กน้อยหรือปัญหาบางอย่าง (ที่ได้รับการแก้ไขแล้วในที่สุด) ฉันไม่ได้มีความรู้สึกแข็งแรงดังนั้นเกี่ยวกับเรื่องนี้ไว้ในสถานการณ์ก่อนหน้า แต่คุณไม่ควรจะต้องผิดหวังกับพวกเขา นอกจากนี้ยังมีผู้ชายที่ดีทำงานที่นั่น
สิ่งที่เกี่ยวกับลาน VPS, BurstNET และแรงขับ VPS? แรกคิดว่าเป็นแรงขับที่วางใน'ถุง'เช่นเดียวกับลานและถ่ายภาพต่อเนื่องไม่ยุติธรรมสำหรับพวกเขา พวกเขาพยายามอย่างน้อยที่จะทำอะไรและยังคงพยายามที่จะแก้ปัญหา'ออก' ฉันจะไม่พูดอะไรเกี่ยวกับลาน -- ฉันคิดว่าทุกอย่างที่มีอยู่แล้วกล่าวว่า ... ฉันสามารถกล่าวว่าฉัน, ส่วนตัว, แน่นอนจะหลีกเลี่ยงพวกเขา สิ่งที่เกี่ยวกับ BurstNet? พวกเขามีเครือข่ายที่ดีที่แน่นอน ... แม้แต่หนึ่งดีมาก ... แต่คำถามคือหากเครือข่ายที่มีค่าที่มีไม่ดี (อย่างน้อยจากประสบการณ์ของผม) การสนับสนุนและไม่ได้ประสิทธิภาพที่ดีที่สุด ฉันไม่คิดอย่างนั้น ในตอนท้ายผมแค่อยากจะบอกว่า -- ผมเขียนบทความสั้น ๆ นี้เพราะผมคิดว่ามันจะเป็นประโยชน์สำหรับผู้ใช้ที่ต้องการงบประมาณ VPS เซิร์ฟเวอร์ ฉันเดิมพันผู้ใช้บางคนอาจมีประสบการณ์ที่แตกต่างกว่าผม ฉันจะโชคร้ายเพียงเขียนขึ้นอยู่กับ'การเดินทางของฉันกับ บริษัท เหล่านั้น
ผมเชื่อว่าคนมีทางเลือกฟรีเพื่ออ่านข้อมูลเพิ่มเติมเกี่ยวกับ บริษัท เหล่านั้นพูดกับคนที่มี VPS อยู่แล้วและคุณสามารถของหลักสูตรการใช้วัสดุที่นี่เพื่อช่วยคุณในการตัดสินใจ มันเป็นของคุณหลังจากทั้งหมด
< <snipped> > ฉันหวังว่าอย่างน้อยส่วนของบทความนี้มีประโยชน์กับคุณหรือทำให้คุณยิ้มได้ ในที่สุดฟอรั่มเว็บที่ผมอ่านผม passive เสมอดังนั้นฉันคิดว่ามันจะดีไปอย่างน้อยบางสิ่งบางอย่างให้กลับ
หากคุณมีคำถามใด ๆ พวกเขาเขียนนี่ผมจะพยายามที่จะตอบคำถามเหล่านี้
Quickweb
Affiliate : <<snipped>>
Non-affiliate
http://www.quickweb.co.nz Clubuptime
Non-affiliate
http://www.clubuptime.com/ YDGH
Affiliate: <<snipped>>
Non-affiliate
http://www.yourdomaingoeshere.com BuyVM
Non-affiliate
http://buyvm.net/ Chicago VPS
Non-affiliate
http://chicagovps.net/ BurstNet
Affiliate <<snipped>>
Non-affiliate
http://www.burst.net/ Thrust VPS
Affiliate <<snipped>>
Non-affiliate
http://www.thrustvps.com/ Yard VPS
Affiliate <<snipped>>
Non-affiliate
http://yardvps.com/ PS. Why am I providing links for services I did not have the best experiences with? Because I believe the choice is completely yours.
End note: Please bear in mind that it is only the first version of this report. It can have some errors and is probably missing some information. I just wanted to share with you what I already have, as I believe that it could help someone.
http://www.webhostingtalk.com/showthread.php?t=1004278 
< กดยุบ (ห้องยกเลิกการใช้งาน)
