Where should I put my server workloads?

The important aspects to consider when deciding between on-premise, public and private cloud servers.

Businesses today face a myriad of conflicting information regarding server deployments and what solution may be most beneficial for their particular requirements. As technology evolves, the options available have equally increased and the modern business can now choose between several Server Workload solutions, which include harnessing the benefits of the cloud.

With many key industry influencers believing that every CIO will decide to be almost entirely public cloud reliant by 2016, it is all too easy to presume that this is the best and only choice for the future. In reality, each business is different and each company will be unique in its IT requirements. To this end, this article explores the important defining differences between on-premise, private and public cloud servers so any business can understand the pro’s & con’s of each & understand which is most likely to fulfil their specific and often demanding IT needs.

To find out where you should deploy your server workloads, simply read the information below (broken down into the 12 key areas we believe any business should consider) before choosing the option which most closely satisfies your requirements for each category. At the end of the article, simply click on the button to reveal the solution best suited to meet your requirements.

  • Virtual Machines

    On-premise: On-premise servers provide easy access to virtual machines, whilst also opening up an extensive array of hardware vendor options enabling a flexible range of solutions. That said, On-premise servers can traditionally be very expensive and as such many businesses mistakenly opt to buy cheap which limits their capacity and future workload growth.

    Public Cloud: Public Cloud servers are both highly scalable as well as ideal for ground up application development. However they can lack implementation and migration support, and often only provide a limited range of OS and hypervisor support. Furthermore, with Public Cloud servers it can be difficult to pinpoint where the data actually resides as it can move between locations, which may not be ideal for all industries. That said, Public Cloud solutions often provide extremely granular billing so your business can understand the exact cost incurred, sometimes even down to the number of minutes consumed.

    Private Cloud: Private Cloud servers provide seamless access to a world of business enhancing enterprise class technologies, with management of the platform typically included by the service provider as standard. This allows for scalability that is paid for as and when resources are eventually required as well as incremental project by project implementation and migrations.

  • Licences

    On-premise: Opting for an On-premise server traditionally opens up a wide variety of licence vendors, however this can equally present expensive on-going requirements where the end customer must manage their licence usage and entitlements or additional costs if the management of licensing governance is outsourced. Licence governance is key but can lead to a complex and ever changing landscape that demands plenty of internal focus, particularly where licencing needs are significant. Additionally, challenges can be faced should transfer of licence assets be necessary in the future.

    Public Cloud: Whilst the Hypervisor is included as standard (with OS traditionally provided by way of an upgrade), there are a large range of service provider licences available although application version support is known to be limited.

    Private Cloud: With the Hypervisor and OS included in the cost of the Virtual Machine, Private Cloud servers provide a company with flexible licence options, including perpetual licences or pay as you grow, which can be ideal for the business that requires an adaptable approach. What’s more as these are managed by the service provider, key internal staff can focus on their core duties, safe in the knowledge that they are covered appropriately and can have access to immediate expert strategic advice when necessary.

  • Change Management

    On-premise: Change management can become an overlooked task for the On-premise server owner. However, selecting an On-premise server is likely to introduce more control, process and governance which, in turn, understandably comes at a cost.

    Public Cloud: All change management within a Public Cloud server is managed by the service provider, although there may still be internal requirements for change management for the OS and company specific applications.

    Private Cloud: Private Cloud servers also offer flexibility here. It is possible for change management to be completely outsourced to the service provider, but in addition, those companies that wish to maintain some control of their own change management can also be accommodated.

  • Capacity & Availability Management

    On-premise: This area of system management is typically an internally handled process for the company with an On-premise server, therefore requiring necessary skillset to implement. It can also mean that the server would benefit from a third party monitoring platform and a completely custom set of definitions from the company’s point of view.

    Public Cloud: Public Cloud service providers provide metrics for Capacity and Availability Management, although they are only platform specific. If a company requires something more comprehensive, a secondary platform will be required with APIs to provide for both sets of statistics.

    Private Cloud: Private Cloud service providers typically provide fully managed Capacity and Availability Management up to the operating system. What’s more with the flexibility of adding management of company applications, businesses can have complete control over the ways in which their systems are managed.

  • Systems Management

    On-premise: It’s fair to say that depth and breadth of skillset is key to systems management, therefore, when working with an On-premise server, the demands upon skilled IT staff can be significant if the system is to be fully and properly managed. What’s more, it’s relatively common for system management to be poorly planned for, which may lead to a catalogue of issues down the line. Finally, if managed effectively it will however ensure a complete understanding of all applications in use & knowledge is retained by internal company staff.

    Public Cloud: Whilst Public Cloud servers include systems management up to the hypervisor, anything beyond this must be internally managed. This can sometimes result in problems for companies with a complex range of applications.

    Private Cloud: Private Cloud servers are SLA driven, feature pay as you go scalability and are ideal for companies where non-core activities require robust management (such as OS support, File and Print, Mail and AD). Furthermore, Systems Management is concisely documented with an agreed patching and maintenance schedule.

  • Incident / Request Management

    On-premise: On-premise servers benefit from a dedicated service desk (either inhouse or outsourced) to manage incidents and requests where documentation and processes need to be strictly adhered to, which can be an obvious challenge for some businesses due to the necessary staffing requirements and / or additional costs. Additionally, On-premise servers may lack trending and reporting functionality that can inhibit continuous improvement of the environment.

    Public Cloud: Whilst Public Cloud servers provide Incident and Request Management for the platform, service providers can be known to be both difficult to receive a timely response from, as well as being frequently consigned to email only. That said, Public Cloud services are known for being particularly stable & robust, therefore it is unlikely that Incident & Request Management will play a significant part in the service delivery.

    Private Cloud: This service is provided as standard by Private Cloud servers and can be completely customised to suit the needs of the company in question. The service can be delivered for all Incidents and Request Management including desktop and external customer support.

  • 3rd Line Support

    On-premise: Support for On-premise deployments can be delivered by either internal IT staff or an external partner – both of which have associated costs. What’s more, if the responsibility falls to an internal IT team, the expense of gaining depth in the range of expertise required to support issues effectively may prove difficult and / or costly.

    Public Cloud: Public Cloud servers provide third line support up to the Hypervisor only with the service typically included within the overall cost.

    Private Cloud: Whilst third line support is provided up to the Hypervisor with options to further support OS and applications, some Private Cloud service providers struggle to provide key application knowledge that is essential for some companies and / or industries, so we advise detailing specific requirements beforehand.

  • Monitoring

    On-premise: The monitoring of an On-premise server demands that a dedicated platform is built around the system, as well as a requirement for internal staff who hold the necessary skills to analyse the data to ensure alerts are not missed and the system can be fine tuned over time.

    Public Cloud: When working with a Public Cloud server it is once again down to the business themselves to provide a monitoring platform, and even then anything under the Hypervisor can’t be seen.

    Private Cloud: A tried, tested and standardised monitoring platform is typically provided as a standard service for any Private Cloud server, which although can’t see under the Hypervisor can be customised to include OS and Applications as well as provide extended functionality including API / Probe configuration.

  • Backup

    On-premise: Backups are undoubtedly an issue for On-premises servers, with a typical failure rate of between 10% - 50% (although this is often down to poor management). Consequently, restores are time consuming and problematic. These issues are often down to businesses ignoring the problem, as it is not seen as a front line service therefore often opting for the least expensive backup options available at the cost of functionality.

    Public Cloud: Backups can be managed by Public Cloud servers, however they do often demand a lot of configuration or an additional product. In addition they are self-service driven, with the provision of an individual portal to allow for the restoration of data.

    Private Cloud: Third line support is available for the Private Cloud server providing additional reassurances for a business. Additionally, Backup can be delivered in a hybrid environment enabling one single methodology and a customisable solution that can suit any data policy – right down to splitting OS, Application and File data for different systems. Furthermore system recovery is client driven and very easy to understand and implement, therefore recovery is quick and seamless.

  • Disaster Recovery

    On-premise: Disaster recovery for the on-premise server requires a custom infrastructure (which must be of a decent quality if the chosen product is to properly handle failovers). Generally, companies have a storage recovery time of between four and 24 hours, which can be harmful for business operations.

    Public Cloud: For the most part public cloud servers don’t provide disaster recovery as standard and this often then demands a second set of virtual machines (which in turn requires plenty of configuration) to be purchased in another location. As such, the responsibility of RTO and RPO remains with the business.

    Private Cloud: Data recovery is simple to use and interwoven in to the designs of a private cloud server. It is SLA driven and is customisable to the storage, Hypervisor or application which helps manage costs. Furthermore it provides quick response and recovery times that can be as rapid as a few seconds for simple tasks, or up to 24 hours for complete storage replication.

  • Connectivity

    On-premise: Connectivity can be a problem, especially if it is not planned correctly. Due to On-premise servers most often running upon consumer grade broadband, such solutions do suffer from a lack of access to Tier 1 internet providers. And, with in-house skills often in short supply, network designs can often fall by the wayside.

    Public Cloud: Public Cloud servers offer Tier 1 carriers for internet and unparalleled scalability, although this must be balanced with difficult to establish dedicated communication lines, as well as firewall configurations which can be costly.

    Private Cloud: The majority of Private Cloud service providers can offer expertise for connectivity infrastructure, however they often fail to provide WAN and core networking knowledge. Of these providers, some have Tier IV DC’s with Tier 1 Carriers but fail to provide cost effective Layer 2 communications, therefore it’s important to select the supplier carefully as there are specialist providers (Panoptics included) who can provide a more complete solution.

  • Exit Management / Novation

    On-premise: When the time comes to migrate to a new platform, On-premise servers traditionally face complications with outmoded standards and, should requirements increase substantially, the company may literally outgrow the capabilities that they can physically house or handle on site, which can be an inhibitor for business growth.

    Public Cloud: Public Cloud servers commonly suffer from workloads that are not particularly portable and an emphasis on the client being responsible for data recovery with little or no support from the service provider.

    Private Cloud: Private Cloud servers offer flexible Exit Management including workloads that are extremely portable as either Hyper-V or VMware. Additionally, Private Cloud terms are negotiable and the service can be delivered as a project. ISO accreditations are also commonly available to boost a client’s capability.

  • This field is for validation purposes and should be left unchanged.

Contact Us

We’re always available to discuss any IT project and would welcome the opportunity to talk it through. Get in touch now and we’re confident we will find the perfect solution for your business.

call: 0203 137 6351 email: hello@panoptics.com

421 New Kings Road, London.

  • This field is for validation purposes and should be left unchanged.