Where’s ‘the cloud’ and should I go hybrid?

Article first published on Spot On Magazine

Today’s start-up and SME businesses are bombarded with an ever evolving array of technical terminology, that many struggle to decipher, let alone understand what it means to their business. Even businesses with CTO’s and technical engineers aren’t always equipped to fully understand new and emerging technology or what hosting set-up would fit their business needs, now and as they grow. Making sure your hosting infrastructure fits your operational needs, as well as commercial restraints can be challenging.

Should my business choose dedicated hardware? Utilise cloud? Or look at a hybrid set-up?

Let’s start this article with a quick jargon buster:

Physical Dedicated Server means physical dedicated hardware hosted in a rack.

Cloud encompasses a few different technologies, but in essence means virtualised compute resources and technology. Cloud technology allows for you to scale resources and gives you the flexibility to quickly add in resources as and when you need them.

Hybrid Infrastructure means a mix of technology such as physical dedicated servers, coupled with cloud to form the overall server-side architecture.

Colocation means leasing rack space in a datacentre facility. This is normally used by businesses that want to retain ownership of their own equipment and ‘colocate’ them in a facility which has resilient power, cooling and fire suppression.

So, which solution is the best fit for your business?

To answer this question a good place to start is asking:

• Do you want to operate an OPEX or CAPEX model? OPEX is the most popular with start-ups and SME’s, as it allows for you to lease hardware and compute resources, without the headache and expense of managing the lifecycle and ongoing maintenance of the equipment. Whereas a CAPEX model allows for you to buy and maintain your own equipment and choose to colocate the equipment in a datacentre facility via a colocation provider. Using a colocation provider allows for you to leverage their purchasing power, as they will already have presence in the facility.

• Are your hosted services uptime critical? If the answer is ‘Yes’, you will want to look at a Highly Available (HA) infrastructure set-up. Simply this means no single point of failure in your hosting typology. For your cloud infrastructure to be HA, you’d want at least two of your core production servers i.e. two web servers and two database servers. If you’re using dedicated hardware, you would need two dedicated servers to allow for failover. These are normally called ‘primary’ and ‘secondary’ or ‘master’ and ‘slave’. If one server goes down, the other server takes over.

• Where is my data physically stored? Is it in the UK? GDPR has put a spotlight on data sovereignty and the importance of ensuring yours and your clients data is properly protected. This starts with understanding where your data is being hosted and stored. If you ask you hosting provider and they reply with ‘in the cloud’ ask where that cloud is physically hosted – London, US, Hong Kong, Mars?

When speaking with a Cloud Infrastructure and Managed Services Provider the conversation should start with a business level assessment of your requirements. Launching straight into the technical discussion can miss out some crucial requirements, which later on can negatively impact your business. For instance, if you run a 24/7 B2C ecommerce business, chances are you’ll need 24/7 emergency response, as a part of your Cloud Managed Service Plan. You’ll also want to consider your options to minimise service disruption. Only having a single web and DB server will be cheaper, but it’s non-resilient. If that single server goes down, your services go down. You don’t have to be tech savvy to ask these questions. It’s about understanding how your hosted services are set-up and asking the question: ‘what would happen if….?’

Client Case Study – Moving from Cloud to Hybrid Infrastructure Model

We worked with a telecoms business to migrate part of their services from cloud to dedicated hardware, which turned their cloud infrastructure model into hybrid infrastructure model. By changing from a purely cloud set-up to hybrid saved them 30% on their monthly infrastructure spend. Here’s how:

The Brief:

We were asked to assess the telecoms company’s cloud infrastructure set-up on Amazon Web Services (AWS). The main aim initially was to see where more resilience could be built it and do a light weight cost analysis.

After a management meeting with the company’s Managing Director, Finance Director and Technical Lead it became clear that this was not just a technical issue, it was also commercial. The company’s monthly spend on AWS had been increasing month-on-month, which quite rightly was getting the attention of the Finance Director, who wanted to see how this could be bought back under control.

Our team worked closely with the client’s internal development and management team, to get a detailed understanding of the services hosted in their AWS environment:

• Our technical team conducted a detailed review of the environment; reviewing the network typology, configuration, security and back up strategy of the 40+ VM’s across multiple regions and Availability Zones (AZ’s).

• From the assessment we delivered a technical ‘health check’ report, which detailed actionable recommendations and insights, for improving the resilience and high availability, improvements to security, best practice user management, increasing performance and better resource utilisation.

Due to the concurrent high-uptime nature of the company’s services, the hosted software had extremely high IOPS activity and disk space usage, which was causing the spikes in expenditure on their existing cloud set-up. Due to the Instance restrictions on AWS, this meant the company had no choice but to regularly upgrade their Instance size, to accommodate their growing resource requirements. This led to an ever increasing infrastructure spend and an inflated cost model, which would continue to grow.

After delivering the initial report it was clear that the company would benefit from migrating to a hybrid infrastructure model. After consultation with the company’s management team, we were ask to propose a hybrid model, with a migration plan for moving the client to the Scholar network.

Solutions Overview:

• We proposed physical dedicated servers for their high intensity database workloads, which would allow for them to utilise dedicated hardware to process and manage the scaling and IOPS intensive element of their systems.

• For their other less IOPS intensive workloads we proposed utilising cloud virtualisation, with tailored VM’s to fit their specific requirements. The fixed instance model on AWS was not the most cost or resource efficient solution, so it was decided that the company would migrate their services to our Scholar VMware Cloud platform to allow for bespoke sizing of the Virtual Machines (VM’s).

• The Backup Strategy also needed to be strengthened to allow for the company to restore from backup in the event of an incident. We proposed an offsite backup to two diverse locations; using Veeam Enterprise Snapshot Backups and Agent Based Scholar Cloud Backup.

• The services were extremely uptime critical, so we integrated a number of safe guards into their hosting topology which included: load balanced and highly available configurations of the VMs, the ability to live migrate VM’s to avoid downtime, as well as 24/7 emergency response for production services.

The end result for the client was a successful migration onto a hybrid infrastructure model, coupled with a 30% monthly cost saving on their overall infrastructure spend.

Conclusion

For many modern businesses their hosted services form an integral part of their business operations. Business Continuity questions, like the ones highlighted in this article, should form part of your decision making process and be questions your Cloud Infrastructure and Managed Service Provider should be happy to help with.

If you can, engage with a Cloud Infrastructure and Managed Service Provide early on in the planning stages. Whether you’re launching a new service, or looking at making changing to an existing set-up, they will be able to give you advice and guidance on your options. This consultation may not be free, but could save you time and money in the long run.

Lastly, don’t be afraid to ask dumb questions – if you don’t know the answer and it’s important to your business, it’s not a dumb question!