Why your network might be a secret killjoy for exciting new applications

The ubiquity of the Internet has allowed the development of a whole bevy of exciting new possibilities. An application can be pulling information across from many different locations and combining it to produce a service that would not have been possible a decade or two ago, or even quite recently.

For example, your order system could be combined with Google Maps to give a visual schematic of a customer's location. But if a significant portion of your services are delivered "from the cloud", there could be something preventing your employees from enjoying quite the experience that was intended: your network. In this feature, we look at the kinds of applications that could be seriously impeded by poor network performance.

As we have argued in two previous features on the infrastructure implications of cloud services and its impact on security, the outsourcing of various types of business software provisions to network-based services has implications for your network.

Although the cloud is the focus of delivery for these services, the services themselves are all about collaboration and mobility, as well as "big data", where trends can be gleaned in greater depth from amalgamating as much data from as many heterogeneous sources as possible.

The mobile worker

In our previous two features, we have already focused quite a lot of attention on the mobile worker.The plethora of smartphones, tablets, and hybrid devices now inhabiting the office worker's pockets and bags means that there has never been a better time to take business activities out on the road.

Supported by carefully selected cloud services, the mobile worker can keep up to date with email via their smartphone, and even take a quick look at reports and presentations that have been sent through. Add a tablet into the mix, and the mobile worker can even work on documents in any spare moment, with hybrid devices providing a good compromise between a traditional notebook and a tablet.

For these portable devices to be most effective, however, the software and documents a worker uses at their desk must be easily accessible when on the move, and as much as possible without conscious effort. This is where the cloud comes in, and with it a greater reliance on the infrastructure and security of your corporate network.

If your mobile workers are making regular use of public cloud services like Gmail and Google Docs, or Dropbox, they will need access to these when within your local area network as well as on the move. This will increase the traffic across your network, and also make its perimeter more permeable.

So you will need to ensure the bandwidth and latency can cope, and that security is both robust enough, but also not so much that it actually prevents the new ways of working from being possible at all.

Virtual desktop access

If you plan to give mobile workers virtual desktop access to their main systems when out and about, or even run a service provision based primarily around virtual desktops and applications, these factors will be even more greatly accentuated. Depending on what your users do on their virtual desktops, each one could require a significant amount of bandwidth, and low latency to ensure a tolerable experience.

Network security will also need to be configured so that the necessary resources can still be accessed by mobile workers. Without these considerations, the benefits of mobile working that have exploded in the last five to ten years will not be fully realised. Your network could easily be the factor that holds things back.

Allied to the explosion in mobile devices is the facility to collaborate. This is built into public cloud services like Google Docs, where multiple people can edit a document at once. Office 365 also provides facilities to share documents for collaboration via Microsoft OneDrive. This provides Google Docs-like web versions of the main Office applications, so that users can edit files without having to have the necessary software installed locally on the device they happen to be using at the time.

Just as in the virtual desktop case, bandwidth will be necessary for a comfortable user experience, if these services are being accessed from within your local area network, although as they are located in the public cloud mobile workers will be using their mobile data connection when on the road.

Latency will be less of an issue, as the AJAX-based system used by Web apps of this type means they function quite a lot like locally stored software after the initial loading time.

The big data revolution

The big data revolution will have a further impact on your network. By its very definition, big data is all about huge amounts of information. The theory goes that using smaller, more selective data sets will limit the scope of the conclusions that can be made. If your data set actually excludes the factors that are determining the behaviour you are analysing, you won't be able to see the necessary patterns.

With big data, in contrast, every possible factor is taken into consideration, and the extraordinary computing power now available is brought into to play to look for patterns. Some companies, such as HP, are now talking about serving brontobytes of data. This isn't an official term, but is generally considered to refer to a number of at least 1,000,000,000,000,000 bytes.

The storage of data of this order of magnitude is just one of the problems it poses. Hard drives are currently at 4GB capacity apiece, so clearly a lot of them would be needed to host this much information. Even with a very sophisticated Fibre Channel fabric infrastructure, very few data centres have the capacity to handle this amount of data.

Facebook, for example, has a three-tiered storage system with capacity for 100 petabytes of data (100,000,000,000,000,000 bytes). Amazon users stream around 40 petatybes of videos a month. The largest data centres in the world store exabytes of data, which is 1,000 times more than a petabyte.

In reality, though, storing this much mission-critical data in one place would be a massive risk, and the data will actually be mirrored and replicated in more than one location. In fact, different sets of data will be hosted in different places, and potentially all over the world. This puts the emphasis as much on the network as it is on the speed of the underlying data centres. To pull all the necessary data together from heterogeneous sources, fast low-latency connections between all of them will be required.

The rise of graph data

Services like Facebook's Open Graph are making a wealth of data about individual behaviour available for analytics, the effect of which is only just beginning to be felt by marketing professionals. Many companies make different types of data available via application programming interfaces (APIs), in particular Google with Maps, Analytics, Translate, Picasa images, and so on, although its weather API was shut over a year ago. For example, the methods for accessing the APIs for Google's Apps Platform can be found here.

Developers are free to combine capabilities from multiple vendors into new services. A lot of very enabling new applications are made possible by integrating data from live sources in this way. For example, a spreadsheet detailing a stock portfolio could be pulling live pricing information from the Google Apps Finance service.

Social media analytics platforms such as Radian6 or Meltwater Buzz can combine information gleaned from the live Twitter stream and articles on blogs with customer data from a customer relationship management (CRM) database, making it possible to tailor a customer's experience around the conversations they are having on Twitter.

For example, a disgruntled customer can be offered solutions to their problems proactively, or the impact of an advertising campaign can be gauged. Throw in sentiment analysis, based on natural language processing (NLP), which guesses the general mood of social media by the words that are being used, and you can obtain a live view of how a brand is faring after a news report or advertisement is aired. IBM's Slamtracker system can show social sentiment towards tennis players as a Grand Slam match progresses.

These examples blend data from one or more external third-party sources with local information. In the Radian6 or Meltwater Buzz CRM examples, the CRM database will probably be on the local network, whereas Twitter information will be coming from an aggregation service such as DataSift, and sentiment analysis services from the software vendor directly.

IBM's Slamtracker blends match information captured live on court with the social media stream and sentiment analysis to create its reading of public mood towards players. Most of this information can be accessed both when on the local area network and externally, so ensuring the connections to all the different data sources have adequate performance is key.

The crucial importance of the network

Whether your company is merely leveraging the benefits to flexible working provided by mobile devices, taking a more collaborative approach to operations, or bringing the full force of Big Data to bear on its business practices, the message is clear.

All of these exciting new applications rely on fast, low-latency networks, both locally and connected to the wider Internet. Without a strong emphasis on providing the best possible networking and Internet connectivity your company can afford, your business could find itself falling behind competitors who have realised the fundamental importance the network.

This post is brought to you by Comcast Business. Follow @comcastbusiness.