If you’re in the process of launching a startup, you’ll know that every penny of investment counts, and that opportunities to establish your reputation can be as unexpected as they are precious. These are precisely the reasons why most startups look to a turnkey dedicated or cloud hosting model for their application or website.
For startups, as for any growing business, a turnkey (or Infrastructure as a Service) hosting model has several big benefits over hosting on-premises. Benefits like facilitating better capital investment decisions, accommodating unexpected or unpredictable demand, and reducing the risk of development setbacks and unplanned downtime. Here’s why.
The typical hosting infrastructure choices for a startup are an on-premises data center, colocation in a third-party data center, dedicated server hosting or public cloud hosting. You may choose to host a private cloud on top of your infrastructure, but this would only apply to very specific types of startup. Here’s a good article about that.
This post intends to illustrate and compare the virtues, for startups, of the ‘turnkey’ dedicated and cloud hosting models, both of which use an entirely third-party infrastructure and represent a monthly expense rather than an upfront investment.
Choosing a hosting infrastructure model and provider
When deciding between and among on-premises, colocation, cloud and dedicated hosting options it’s important to consider the quality of the underlying infrastructure that each option provides. Remember that the quality of any hosting infrastructure depends not only on the specifications of the computing and storage resources you buy, but also on the quality of the hardware and the network and power infrastructure behind it. Most of all, it depends on the level of redundancy in the infrastructure.
Redundancy means replicating every element of an infrastructure twice, with no common element that could fail without a backup kicking in. Redundancy is what keeps your server running even when failures occur. It’s what stops your servers going down when you’re doing an important presentation, or when your influential first customers are trying it out. Network and power redundancy is how premium hosting providers like iWeb offer a 100% network and power uptime SLA and, most importantly, how we have met this SLA every day since 2008 in spite of isolated failures in our infrastructure. We also offer 100% hardware uptime SLA if you deploy a hosting infrastructure with redundant hardware.
The need to achieve power and network redundancy is the main reason why hosting in an on-premises data center is not cost effective. Building redundancy into an infrastructure is an expensive business, involving vast sums of capital investment and expenses – at least $300,000 per year in networking material purchases and rentals, before you begin to account for expertise, server hardware, redundant power sources and server cooling. For most businesses, this not an attractive way to use capital. For startups, when capital is especially precious, capital expenditure on a data center is usually very low on the list of priorities.
A similar economic argument holds true for human resources, like certified solutions architects and system administrators, which are also available from a good turnkey hosting provider. For a startup it’s unlikely that full time specialist system administration will provide the necessary return on the combined time and money invested in their recruitment. Far better, perhaps, to have this specialist expertise on hand for only the hours of administration that are needed, as sporadic and diverse, as they may be. Once again the ‘on demand’ nature of a turnkey hosting provider becomes a key benefit. It also means that the competency and reputation of your hosting partner becomes extremely important in your choice of partner.
Finally, the flexibility afforded by turnkey hosting affords startups greater scalability, which itself makes the model very cost effective, especially in the short to medium term. A startup could start with a small cloud server for development and testing, scale that server as demand grows or spikes, scale out to a hybrid dedicated-cloud environment then add redundancy with a high-availability server cluster when critical mass is achieved (see below for more details on this kind of scaling). Cloud servers in particular are useful for minimising expenditure without limiting the amount of traffic you can handle, perhaps in response to a soft launch, beta test or a great piece of media coverage that sends a burst of traffic to your site. That’s because cloud servers can scale elastically according to demand, and you only pay for the resources you use for the time you use them, and for the actual traffic to your application or website.
A case for the alternatives: on-premises and colocation
The cost models for on-premises and data center colocation are very different from the recurring monthly cost of Infrastructure as a Service.
Building a data center or installing your own hardware in a colocation data center allows you to invest upfront in material and amortize that investment. In the case of colocation, you rent the data center space with redundant network and power, and you buy, install and manage the server hardware yourself. If you have the skills in your company to set up and manage the hardware, this could work out cheaper over a long enough period of time. On the other hand, if your hardware fails you will need to replace it yourself.
In order to make economic sense, you may need to keep and amortize the hardware over several years, a period of grace that is not guaranteed for a startup, and something that would prevent you from cancelling or migrating to a newer server without penalty (providers like iWeb have no minimum commitment – you can easily close/migrate servers after only a month).
Hybrid: the best of two worlds?
If you website or application is your business/service, it can’t afford to go down. But maybe your internal services, the ones that you and your colleagues access, are not quite as crucial. If this is the case, you could set up a small data center in your office for your internal network, perhaps also for development and testing, or for any other services or data that do not need to be available to your public audiences 24/7.
To take this combined on-and-off premises further, a more advanced consideration would be a hybrid hosting model like a Microsoft Private Cloud, which provides unified management across cloud, dedicated and on-premises infrastructure, allowing for easy migration and scaling between hosting infrastructures.
Choosing between cloud and dedicated options
Assuming you are convinced that a turnkey hosting solution is the way to go, now you need to choose among the many services available. First consider the factors discussed above – the quality of the infrastructure, level of redundancy, the professional services available and the network and power SLA.
Once you have made a shortlist of reputable, reliable and skilled providers, it’s time to select your compute and storage resources. This involves essentially deploying a dedicated or cloud server, or a high availability server cluster designed to guarantee server uptime and/or make data available to several servers/applications/websites.
Dedicated servers offer higher specifications of computing resources like CPU, RAM and a wider range of storage options including sizes and types of hard drive and RAID options for data security. As it’s name implies, with a dedicated server 100% of the resources are dedicated to you and your server is completely private.
Cloud servers offer a lower entry price with lower levels of computing and storage resources. They also offer lower prices in general, as they use virtualization to pool and allocate resources to many users in a way that is more efficient, and therefore cheaper. The flip side of this less stable performance, although new technology like iWeb Cloud Servers do offer the same stability as dedicated servers.
Cloud servers also offer features designed to make managing your server easier. These typically include fast deployment (a few minutes), the ability to clone your server and the ability to scale your server (add more computing and storage resources) quickly without migrating server. An API can also be used to make cloud servers scale automatically in demand to workload, as well as to provide easier data connectivity over the web. All of these features also make cloud servers conducive to pay-per-use pricing, where you pay for the time your server is live and the traffic between the server and it’s clients.
These different features tend to make cloud servers more convenient for smaller websites, for development of applications, for any application or website with highly variable or spiky workloads. Dedicated servers are useful for hosting several websites, high traffic/workload websites, large databases, private cloud environments or enterprise applications and file sharing.
For more detailed information about these options, chat live with a sales assistant on iWeb.com who can provide further information free of charge and with no obligation to buy.
Startup hosting packages by iWeb
Thinking of launching a startup? Chat Live with a sales assistant on iWeb.com for some free advice on choosing a server. We also currently offer the following discounted bundles designed for a smooth transition through the gears to a successful future. And whatever you choose to do, good luck with your endeavour. We’ve been there too.
These prices are correct at the time of publishing. Please chat live on iWeb.com for the latest prices and options.
|Phase||Ideation & Creation||Development & Launch||Growth & Maturity|
iWeb Cloud Server
|Hybrid Cloud Hosting
iWeb Cloud Server
|High Availability Cluster
2 Web Servers
|iWeb Cloud Backup 100GB
|Dedicated Backup Server
|Included Services||Support On Demand||Support On Demand
|Support On Demand
Priority Expert Support
No comments yet.