Behind the scene's of a hosted desktop service

From simply looking at the desktop or even testing a hosted desktop for a period of time will not provide you with an insight into the technology behind the scene that ultimately depicts whether the service will continue to be quick and reliable and whether you can be sure your businesses data is stored and backed up correctly.

Although different Hosted Desktop provides may use the same technology e.g. Terminal Services, the platform that this runs on can be vastly different, and it’s this platform that you are trusting one of the most valuable parts of your business with, your financial and company data.

Brutal Fact: Even though marketing material might suggest otherwise, your “cloud computing” or “Hosted Service” is in fact a Microsoft network hosted at another location. AND the kind of issues that can affect a typical in-house network can also affect a hosted network.  There will be lots of talk about this being 99.999% reliable etc BUT you need to face the fact that this is, at the end of the day, a Microsoft IT network and as such will at some point have issues. With that said, moving to a Hosted Desktop service does offer more reliability than a typical network because they are normally hosted on enterprise level, more reliable equipment with more appropriate management of the environment taking place. Please don’t envisage this is a magic wand, the ongoing support and businesses you chooses to host your desktop is as important, if not more important, than the different in monthly price you may pay.

There is a large sliding scale of investment that can be made to host a company’s network. Below is an outline of some of the most common ways companies provide a hosted desktop service.

Using VDI (Virtual Desktop Infrastructure)

VID is a product produced b VMware. This is similar to simply picking up and moving your PC and network into a data centre then connecting to it remotely. In reality the PC is a virtual PC and runs on servers much more powerful than your desktop PC.  At a basic level this is just like connecting to your office desktop PC remotely and depending on the configuration you may over time experiancve the same kind of performance type issues as more applications and patches are installed. Some providers remove this issue by “locking down” the image to avoid users making changes however VID is often used because of its flexibility.

Using Terminal servers

The more widely used Terminal Services offers the most reliable, saleable and mainstream solution for presenting hosted desktops. This technology has been developed over the past 15 years and is now the standard solution for providing remote access. The terminal server is highly efficient with its resources and because of this many users can work off a single server. The processing power of a server is now accessible to the users working via the terminal server and because of this the speed people are able to access applications like word, outlook etc is much faster than a standard desktop PC.  The main downside to terminals services as with VDI, is that sound and video performance is often poor. Recent improvements with Microsoft Server 2008 have improved on this however video editing, CAD and other similar requirements may still be best served from a Desktop PC. This however doesn’t stop the rest of the organisation using a Hosted Desktop, it simply mean that this small group of users needs to access servers via a VPN and separate comms link.

Using Citrix

Citrix has been around almost as long as Terminal Services. Citrix is actually a bolt-on to Microsoft Terminal services and requires the initial terminals services license and a Citrix license to run. Citrix has historically offered IT manager’s simpler methods for delivering applications to a desktop than Terminal Service. In the most recent version of Windows 2008, Microsoft has included many of these features. Citrix is still widely used in enterprise level organisation (1000+ users) because of the simplified back-office management. The main reason most smaller organisations would look at Citrix is because of its greatly improved handling of video and sound. In terminal services a video may be jerky and sound may stutter, Citrix reduces this lag by using an improved connection client. This connection client or software improves the speed and reduces and jerkiness of a video.

Back-office Infrastructure

Sharing physical servers

An IT support company can reduce the costs of their service by hosting several companies on a fixed set of servers. This can range from sharing the Terminal Server (the actual server you log into to access your desktop), to sharing some of the less obvious background servers like a file server, Exchange email server or backup server. Although reducing the cost, several organisations sharing resources in this way is a sure fire method for disaster. Each organisation uses different software, printers, network configuration, works different hours and expects different levels of availability. When an issue arises with a customer it will then have a direct affect with all the other staff of all the other company’s using that hosted desktop service.

Using a SAN based infrastructure

The best method of hosting a desktop is using high performance centralised storage. These SAN’s (Storage Area Networks) off ultra reliable and highly redundant storage. This forms the underlying platform for the servers to connect to. In the event of a physical server failing the data is maintained on the san and can be accessible in a number of minutes to hours.

Virtualisation with a SAN Based Infrastructure

Virtualisation (often using VMWare) merges the performance and reliability of SAN technology with the fault tolerant and high availability features of VMware. In the event of a physical server failing (which would normally result in hours of no service) the VMware technology can ensure there is no loss of service. In addition to this Fault tolerance, VMWare provides the facility to take Snapshots (entire copies of the wholes server from a point in time) and copy these to another server or alternative location. In the event of the entire data centre becoming unviable the server can be brought online within hours (usually this would take days).

Virtualisation with a SAN based infrastructure and Synthetic Transaction monitoring

Every company relies on specific key systems and applications that are vital to its day to day business and long-term success.

The Back Office Infrastructure and applications that serve personnel and customers, such as Email servers, SQL servers, CRM systems and web based services are required to be available twenty four hours a day, seven days a week. Outages can have disastrous effects on an organisations ability to conduct business, its reputation and ultimately its profit margins. However ensuring these mission critical services are available every hour in the day is an involved and intricate responsibility.

Synthetic Transaction level monitoring makes sure mission critical services such as email, are always available, responding when required and achieving the highest levels of performance and reliability necessary to deliver the best possible service to your staff.

Synthetic Transaction monitoring provides a best-in-class monitoring service for the software environment as well as the hardware platform.

SNMP (Simple Network Monitoring Protocol) is the industry standard monitoring solution has been utilised for over 20 years and connects directly to equipment being monitored e.g. a Network Switch and reports back the basic information being gathered by the device. The SNMP monitoring information will typically include straightforward information like if the device is turned on, how busy is a servers CPU or how much space is left on a hard drive.

Proactively monitoring a network with SNMP will help to highlight a busy CPU or avoid downtime by alerting an IT Support company that a hard drive has only 5% free storage left, however the service is unable to recognise software such as operating systems or applications.

 

On average 96% of faults are caused by software related issues. Because of this, STM (Synthetic Transaction Monitoring) was developed to offer a more sophisticated monitoring capability that not only reports on the SNMP information but also recognises the operating system and applications e.g. Windows Server, exchange etc, and then compares the current configuration and performance against a preconfigured “application management pack”. 

 

The Application Management Packs (the modules used for specific applications) are designed by the development teams within Microsoft who developed the various packages i.e. Server 2003/8, Exchange, SQL etc. The in-depth experience and knowledge gained from the Microsoft developers is entered into the Management Pack to provide an ideal system and application configuration and a benchmark for performance for each element of the application. 

 

“SNMP works at the basic hardware level whereas STM has the ability to delve into the application layers and its here at this multifaceted layer that the potential to cause the most problems is apparent”.

 

STM will assess the entire environment (both hardware and software) then highlight any settings or configurations that are not in-line with Microsoft’s best practice for that application. STM then prepares a report for an engineer which outlines the cause of the issue and recommends the best course of action for resolving it.

 

Using these more advanced monitoring tools are often the difference between accessing a service which is fast and reliable and constantly calling the Hosted desktop provider to log reoccurring or intermittent performance faults which can often be very difficult to diagnose without a high level proactive monitoring tool.

 

 

The Desktop Technology