Posts from — January 2011
“Some people tend to go for freebies, but would you take free eggs if they were offered on the parking lot of a supermarket? I know I wouldn’t”.
What is affordable anyway? At first, you could think it solely depends on the price. In most cases you end up realizing it’s rather far away from that. Taking the cheapest hosting service on internet is not far away from buying the cheapest house or a car available. I know you wouldn’t do that so read on and i’ll explain why you shouldn’t get the cheapest hosting service provider either.
There are many hosting companies out there offering web hosting for ridiculous prices. They offer everything unlimited for only 5.95 per month. At the same time they talk about global sales, special trained staff, fiberoptics, fully redundant cisco networks and so on but how can they get all the money for building the infrastructure, equipment and training then? Not from hosting. You can see it from their price tag.
While these companies render themselves economically unwit by the promises that are proved to be impossible, they also reaveal their true knowledge of technology by claiming to offer unlimited anything. The truth is, nothing is unlimited, except outer space, and this we are still unsure of. So unless we prove the space to continue forever and get it captured inside a hard disk cover, let’s just assume there is no such thing as unlimited. Especially when it comes to bandwidth and disk space.
The major concern when it comes to choosing a web host is the speed. Not the bandwidth or the disk space as you can always get more. But when it comes to server speed, or responsiveness, that is something that you cannot get with money. It’s just what it is and usually gets worse by the time when the server fills up. If your host offers unlimited features with a low price, you can be sure that the servers need to be filled with a lot of customers to get profit from the server. And that means slow speeds which will never get any better.
The next concern is what you can do with the service. Many companies like to play safe and simply install everyhing into their servers and offer that as a single package. But what happens on a computer that runs ASP, ASP.NET, PHP, Perl, Python, mySQL, MSSQL, SSI and so on? In worst cases, nothing. In a milder variation of the above, something but not what they truly intended. When they get it all working, the server will have so many workarounds configured and security holes created so that it should be drawn out of production to be repaired. Unfortunately on some companies this doesn’t apply and instead of being concerned about clients security, they’re more concerned to get everything somewhat working so they can write all the fancy features in a single ad on their site.
The time goes by, they apply some updates and nothing works. They apply some more funky configurations and get the computer running once again. This game resembles a lot of the spring-games I used to play when I was still a kid, trying to get the waterflow from melting snow going where I wanted. It’s not like you tell the flow where to go but where not to. No matter how hard I tried, the next morning the water was going where it wanted. The problem wasn’t with the tools or the technique, but the original idea of getting the water go forever where it wouldn’t go. This is very close to the idea of having all the technologies running on a single box. Cool, but doesn’t work.
When you get your site hosted on a server with the above configurations, you’ll force your webmaster coding in a broken environment. Many times with old drivers because the new ones caused problems. The code that the webmaster codes on the server is forcibly filled with workarounds similar to the ones in the operating system configuration. And when the time comes, that the hosting service provider discontinues their services because they were found to be impossible, you will need to switch hosts. For your surprise, the code won’t work on the new server. Why? Because the new server doesn’t have all those “special” configurations and “hot ideas” in the operating system.
At this point you will need to contact your webmaster to get things working. And pray the god that you can contact the same webmaster you previously used as he hopefully remembers even something about the workarounds in the code and why they were implemented. Getting a new webmaster would probably lead to rewriting the whole code. At least it should be cheaper that way. Sounds pricey? It is, especially when you take into account how much time and money were used with the initial code while coding in all the workarounds we’re about to trash.
So when choosing an affordable host, you should search for a host with reasonable prices compared to the services they offer. This way they can afford keeping all the technology working top notch. This way your environment will be stable and working and the development on the server should be as fast and cheap as possible. Do not believe in marketing gimmicks like unlimited mailboxes, disk space, bandwidth or anything. It’s not even possible and it only proves that the company is driven by marketing people biased to improving their own economy. Not the customer’s. When it feels too good to be true, then it is exactly that.
January 31, 2011 No Comments
Hardware failures, software failures or user error can all cause system downtime. Downtime causes customer dissatisfaction, additional customer support, lost revenue and impacts your reputation. When your web site goes down, you’re losing potential customers and risking existing ones.
The problem is, the majority of web hosting companies have no redundancy in place, with one server handling all requests. A hard disk failure, power supply failure or software issue will bring all websites on that server off line. It may take hours to repair the problem, leaving customers stranded. Carefully read the fine print when choosing a web hosting company. Most provide uptime guarantees that only cover network uptime, not service availability. That way, They guarantee this because as long as the network is functioning, regardless of whether or notif your web site is up, the uptime guarantee is being met.
High Availability refers to an architecture designed for very high uptime. A High Availability solution greatly reduces the risk of downtime by integrating spare devices that will take the place of any resource that becomes unavailable. A High Availability infrastructure has redundancy starting at the network level all the way to the server level.
A High Availability network will connect to multiple backbone carriers and have fully redundant routers /and switches. A provider, router or switch failure should not affect your website. Power systems should be fully redundant including auto transfer to a backup generator in the event thatof grid power failures. Power distribution units should be redundant and never overloaded.
Temperature and humidity should be controlled by multiple air conditioning units, with enough additional capacity to maintain the environment in the event of an AC unit failure.
Servers designed for High Availability include dual power supplies and a RAID (redundant array of inexpensive disks (RAID).) If a power supply fails, a backup will instantly take its place. The failed power supply can be replaced without bringing the server down. RAID refers to using multiple hard disks which appear as one. If one of the hard drives in a RAID fails, no data is lost and only a small performance penalty occurs while the array is rebuilt using a spare hard disk.
Redundant power supplies and RAID are key components in a High Availability infrastructure, although it can be taken a step further. Clustering is a term you may have heard before. Fail over is a type of clustering that dramatically increases service availability. In a fail over cluster, two servers are linked together. One is actively handling requests, while the other stands by. If the primary server should fail, the secondary server will detect this and automatically take over. The switch should take place within seconds, with little noticeable effect to your sites’ visitors. Fail over clustering can protect against total server failure, including file system corruption or critical hardware failure (such as the processor or motherboard.) It can take hours to re-build or provision a new server after a failure, clustering eliminates this possibility.
Monitoring the state of a High Availability infrastructure is equally important as the redundant hardware itself. The network, server and application layer need to be closely monitored on a continual, 24×7 basis. Technicians trained in High Availability need to be available around the clock to instantly respond to any issues that arise. Monitoring should be performed from an off site and off network location, to ensure a true perspective on service availability.
A hosting environment designed for maximum uptime includes redundant network, power, cooling, data storage, server hardware and proactive monitoring. If you have a mission critical web site that you depend on, make sure your next host has an infrastructure designed for High Availability.
HostSG Cloud Sites offer high availability and redundancy based on Cloud Computing Technology with multi-tier network. Apart from this, HostSG also have extensive network connection via Singtel, NTT Communication and PCCW. This ensures our clients fastest connection to Asia and the world.
January 17, 2011 No Comments
At nearly all webhosting companies websites now a days, you will see the two terms “Cloud Hosting” and “Dedicated Hosting”. What is the difference between these? Well its fairly simple.
On a Shared Cloud Hosting network you have your website placed on a computer (The webserver), that has other websites on it - anywhere up to 350 or more. This has a number of down sides, such as being slower, less secure, as well as giving you less access to edit things on the server, such as installing & updating the system software. One the plus side however, it is much cheaper then going for a dedicated server, as well as being easier to administer.
On a Dedicated Server, you have the entire machine to yourself. This is fast, secure and you have full control. You can install any program your heart desires and update programs as often as you wish. This is a good choice for sites that get a lot of visitors, use a lot of system resources (through the use of CGI scripts), if you are holding very important data, or if you need special software that is not available on virtual servers.
- To sum it up, you need a dedicated server if ::
- Your site gets so many hits your Virtual Server account cannot handle it.
- You want to host many sites on the one server.
- You need a feature that is not offered on a Virtual Server.
- You require a high level of security.
January 4, 2011 No Comments