Humphris Hall is a Managed Service Provider (MSP) with customers in the financial sector such as investment banks and hedge funds. Traditionally Humphris Hall concerned itself with run of the mill IT services such as email and desktop services, however it has grown its expertise in general infrastructure services and as a result Humphris Hall set up their own cloud hosted platforms (for email, data storage, line of business applications) colocated in the Datanet AHF data centre and delivered their managed services from here. This model required quite a high CAPEX investment in hardware and, after 5 years, when the hardware was due a refresh, there was a decision to be made, either invest another big chunk in new hardware or save on the CAPEX and go for public cloud, they chose AWS Cloud.
One of Humphris Hall’s clients, having moved all their compute to AWS, has challenged Humphris Hall to a cost saving exercise. They are currently spending considerable annual sums for their AWS solution and would like to reduce this overhead. Initially AWS may seem like a cost-effective solution but, as it is so flexible and scalable, it is also very easy to add more compute resource here or a widget there and as each of these elements are chargeable, before you know it, the costs have spiralled out of control. As Humphris Hall performed their analysis of the costs involved, they could also see that a substantial proportion of the costs were attributed to the storage element in the infrastructure solution and here is the catch: according to James Humphris, MD at Humphris Hall, in simple terms, the performance of the storage that you take with AWS correlates directly to the volume of storage. So, in layman terms, a data store of 2TB will have twice the performance of a data storage of 1TB. This very fact has led IT managers to over-egg the amount of storage they take to ensure that the performance meets the criteria of their IO needs. This would not be the case in a traditional local private data centre where we would optimise all volumes of data.
The client replicates data between different AWS Regions (different parts of the data centre or different data centres) and Humphris Hall noted that AWS charge a substantial fee for data transfer between the two SQL instances, we are talking about large amounts of money just to replicate your own data, again something we would not do within the Datanet data centre regions (Fleet and London).
So, there are many benefits in moving to AWS such as the flexibility, per second billing, endless resource, ease of scalability but it is a misconception that hosting in these very large public clouds is cheaper than the local data centre model. You also lose the personal touch and skills of the engineers at a local data centre and if something goes wrong you have a very limited options, Nobody Can Hear You Scream in the AWS/Azure darkness, you don’t even know where your data is! You have to put a great deal of trust in AWS and have deep pockets!! In summary, the long term costs associated with large public cloud (AWS, Azure) can be much more than your local private cloud data centre provider such as Datanet. For more info or to visit our London Edge data centre please visit www.datanet.co.uk or call 01252 810010 and ask to speak with Jeremy, Lucy or Conleth, yes real names.
01252 810010
Secure Sustainable Hosting & Networks
