About this site


Guess Who Am I (www.guesswhoami.com) and Fangcaoxin (www.fangcaoxin.com) have lived their lives in different hosting providers. This page is about the history of how they have changed over the years and also reviews for different hosting providers.

Guesswhoami.com was originally registered on November 2005 and fangcaoxin.com was registered on August 2008.

Different providers

DreamHost era

Both domains started their lives on DreamHost, moved from shared hosting to the Private Server, got upgraded many times and also took some of the optimizations into use (W3 Total Cache and XCache for example). In the end (just before moving away from DreamHost) they were also running Amazon CloudFront as their CDN.

DreamHost is very big hosting provider and the quality varies a bit from time to time - sometimes you get extremely fast response time from support and sometimes you need to wait for answer quite a long time even for very simple issue. Generally DreamHost has tried to automate everything possible but quite often you hit to the system limitations or errors where you need help from support. DreamHost has had problems in the past regarding their growth and also being one of the biggest also has side effects in the form of attacks. Some of the technological choices in the past for filesystems or networking solutions have caused problems while DreamHost has been able to find pure bugs in Cisco’s networking equipment as well.

Overall, DreamHost is something I would recommend especially if looking cheap and still good hosting. With DreamHost it is possible to scale up quite much because after shared hosting gets small, it is possible to switch into their Private Server offering which is basically one sort of VPS.

Global reach for the blogs

Well, we have relatives in China who enjoy our blogs (mainly they enjoy Fangfang’s blog as I am rather boring and do not post as often as my wife) and also photos about our life in Finland. Lots of photos online will usually result some kind of performance problem and eventually you end up looking for a CDN.


I was looking for faster CDN, which would have better presence (especially in Asia) as well. We had started from CacheFly, switched to Amazon CloudFront and I had tested also CloudFiles (when they were on top of LimeLight still) but nothing looked exactly like I wanted. But then I came across with quite interesting option, there would be possibility to purchase Akamai through VPS.NET and as everyone know, Akamai is the king of the CDNs. I signed up and noticed how much better Akamai is but what I noticed at the same time was how good was VPS.NET experience. Quite refreshing in today’s Internet so I tested a little bit to see how good they would be as a VPS.

And results was that I moved our blogs to VPS.NET. Soon we were running one VPS hosting our blogs and another running our databases (powered with Fusion-io power). And our blogs really started to fly, we basically cut our load times to half. Database server using Fusion-io was performing very well and standard VPS was also up to my expectations and performed well compared to the shared hosting. One very good side of VPS.NET is that their London cloud is very well connected to London Internet Exchange and thus routing is extremely good.

VPS.NET is a new company on the VPS providing business. They have some interesting ideas and quite good pricing although not the cheapest in the market. One of the major problems with VPS.NET during 2010-2011 was the stability. The SAN systems failed numerous times and also otherwise you were able to spot that they have just begun their operations and stability isn’t just there yet. They are using (still when writing this in the end of 2011) software based SAN to minimize their costs and while those solutions can be extremely good, one might raise a question whether it is sufficient for the VPS operations. Very often you experienced really bad performance when the SAN was recovering from disk failure or some other problem. This type of problem shouldn’t be visible for users normally but in this kind of environment and with software based SAN, it was visible. It was in fact so much visible that our blogs went basically offline every time the SAN had some problems or it was just syncing the data. When the customer base started to grow, the standard performance from SAN systems didn’t reach high performance any longer and every time there was a problem with SAN, the server basically fell down to it’s knees and wasn’t able to perform any simple task.

The support with VPS.NET was very good in the beginning but it quite rapidly started to shift towards standard canned answers where you were spotting that customer actually knew more about Linux, VPS and SANs than their support. Growth typically doesn’t come easy and with VPS.NET you were able to spot that the growth caused huge problems to their daily operations.

Then VPS.NET lost their agreement with Akamai and they switched to Level3 CDN with lots of promises how they could have much better relationship and integration with Level3 compared to Akamai. Akamai was the big selling point for me and Level3 with VPS.NET configuration isn’t anywhere near Akamai even if you would enable Level3’s Asian nodes (which by default aren’t included).

This was the phase when we entered into the constant “everything will be improved within couple of months, new features will arrive within couple of months” time. VPS.NET developed a new method to over-promise a lot and apparently hit the resourcing problem very badly - I guess they were trying to keep the customers while most of the customers were thinking about leaving at that time due to the constant performance and downtime problems. It was obvious that VPS.NET couldn’t keep their promises as so much much over-promise in the air.

It started to look that VPS.NET is only good for sites where the uptime isn’t important and it doesn’t matter if you’re down for over 24 hours and you frequently encounter small hiccups. But I didn’t want that kind of problems, most of the time the downtime happened during my traveling and created additional stress. I developed some interesting methods to guarantee uptime by utilizing DNS based failover methods and special monitoring to catch the filesystem problems (during SAN synchronization, the filesystem typically went into read-only mode, which wasn’t easy to spot without proper monitoring). But it started to get laborious, sometimes I had half of my nodes down waiting SAN synchronization to finish or some other problems.

Eventually, I got tired about VPS.NET problems, especially since nothing got improved at all. I started to look someone who would be having a bit more stability and longer history in this business. I came back to old discovery to check how they compare and noticed that situation had changed - since major portion of our traffic is handled by Akamai these days, it means that we don’t need actually so much bandwidth on the server side.

Good things in VPS.NET is quite wide-range of services offered through partners - one being KSplice, which today isn’t any longer available for new customers and is only available for old customers - in VPS.NET case, you would be able to utilize VPS.NET relationship and get KSplice for yourself (VPS.NET is an old customer of KSplice and thus not affected by Oracle’s decision to stop accepting new customers).

One of the problems is that these services often aren’t very well integrated and you do not get access to real interface of the service provider so you are stuck with the lesser experience than buying it directly from the service provider. This is especially true for KSplice and Server Density where their own control panels will offer much more functionality than what is available through VPS.NET.

VPS.NET’s biggest problem is basically SAN, even the newest SAN isn’t completely fixing it. They claim it is the performance leader in the world but it is nowhere near the hardware-based SANs. Hopefully the stability is now in place but I was disappointed about the read performance. One of the good things with VPS.NET is the good range of locations, they have cloud even in Japan (finally, after they promised it for the next month, it took over half year actually).

They also claim that their approach isn’t to oversell but still their VPS is not performing as well as some other competing clouds with less resources. Maybe they are overselling or maybe it is just the configuration they are using. It will be enough for basic web site but it might not be filling the requirement for high performance.

With VPS.NET, it is best not to raise the expectation too high. Also, it would be very good idea to know UNIX inside out so you can troubleshoot most of the problem by yourself. But if you do know UNIX and you are capable of managing your own server, it might be good choice. But still I would be careful with the SAN performance especially if you are looking for some high performing I/O.

Rackspace era

By using less bandwidth, it became possible to move to someone who is traditionally considered expensive due to the charging bandwidth. For example Rackspace, which has utility billing model in use (pay only for what you use). So, moving from VPS.NET to Rackspace would bring us their Fanatical support, their long history of being in business and also their scale (they are one of the biggest) and at the same time it actually would lower our hosting bill.

So, today, both of our blogs are hosted with Rackspace and everything has been just smooth and stable. And if we have needed anything, the Fanatical support has been up to it’s reputation each time.

One of the good things with Rackspace is their size, it also shows in different kind of promotions through their partners. Very often you might find very good discounts on excellent other services provided to you because you are Rackspace customer - whether it will help you or not, depends on the needs.

Rackspace has been extremely stable, I have experienced zero problems with the servers and only problems have been short network and load balancer related troubles which have been dealt extremely fast.

Different blog platforms


Originally I started blogging with static HTML files, then I took b2/cafelog into use which then turned into WordPress. WordPress is very good blogging tool which has become a content management system. And there the problem lies, WordPress is a bit too big today and quite complex. It is very difficult to optimize it perfectly and you will need a lot of time and effort to make it fully optimized. I know that there are a lot of different plugins to create the needed missing functionality for minimizing HTML, CSS and JavaScript but those don’t always work just like you would prefer.

Time for change

Some time ago, I started to look for alternatives just like my brother did. He has described how he changed to OctoPress. I tried OctoPress but it isn’t exactly what I was looking for, the usage of Markdown is good but the whole concept isn’t optimized and adding that into the concept will require some customization. So, then the next option was to actually go for something, which is optimized from the beginning. And here comes Alula, which is created by my brother Mikko.


Alula is a new framework to create optimized blogs. It is based on the idea of utilizing static files, so it is automatically optimized for cacheability and CDN. The world has changed a lot from the beginning of blogs and nowadays there is great need for good cacheability and natural support for CDN. WordPress is dynamic by nature and it creates problems when you try to optimize the performance. Alula is next generation blogging framework to tackle these challenges.


Why work so much just for two personal blogs? One reason is to make everyone happy, Fangfang can update her blog more often and more easily when it is fast and responsive. We will also get (hopefully) more visitors because our blogs are faster. And everyone in the family are able to see what we are doing and they do not have to wait until they get bored to see our newest photos.

And this is an interesting hobby …