Show All
Top 5
The glitterati of the tech world are once again descending on a chilly Las Vegas for the annual Consumer Electronics Show. Over the next few days, reports of the latest smartphones, tablets and netbooks will flood RSS feeds and whet the industry’s appetite for the latest must have gadgets. Of course, most of these new devices are geared towards one thing – driving data demand.
From the latest Android tablets to 4G phones, the speculation as to what will be displayed is rampant. What’s common to many of these devices is the focus on media consumption and sharing, specifically video. Whether in the form of movie downloads, video conferencing or online gaming, video’s impact on the tech industry is in many respects still in its infancy. We have yet to see the true impact of ubiquitous video. In a previous post, I highlighted how Netflix’s Instant streaming service now accounts for over 20% of North America’s traffic, even though only 2% of Netflix customers subscribe to this service.
Yet as video’s impact continues to be felt on our global networks, many within the industry are peering into the future to see what lies beyond. What’s the next killer app? This is a difficult question to answer. There are few who could have predicted video’s impact and the subsequent bandwidth explosion. The speculation as to what comes next is varied. A number of analysts point towards telemedicine, others remote learning and some the smart grid. While all of these are certainly contenders, I believe that work is the leading contender.
At a recent conference, Gigaom’s Om Malik explored the future of work, specifically the shift from the fixed office to a mobile environment. Malik believes the development of the world’s broadband infrastructure has given rise to a greater state of connectedness that enables us to work from anywhere. In many respects this is true. I’m currently posting this blog from a train. Several years ago this would not have been possible. Malik calls this shift from the fixed office and the corporate infrastructure ‘the human cloud,’ the intersection between web and work.
This is a change that I openly welcome and one that I believe companies need to embrace if they are to remain relevant and successful in a globally fierce marketplace. What I find interesting about the concept of work as a killer app is that it’s not really an app. Work pulls together all the killer apps of the previous five years under one enormous bandwidth-demanding umbrella. Video conferencing, file sharing, VoIP, etc, the list of apps that is required to make remote working possible is potentially enormous. Yet this is a shift that has to happen.
From a cultural, business and environmental perspective, the shift to the human cloud is one that is critical. Businesses cannot continue as they have for the past decade. What we need now is the networking infrastructure to make this happen. Video helped pave the way, but now we need to ensure that the network exists to make a true human cloud possible.
Are you part of the human cloud? If so, how has this impacted upon your work? Does the network limit you? How do we need to develop? I appreciate your feedback.
Great post – and its true! Work is the next bandwidth hog – but only because the cloud is going to get misrepresented by the quick sale merchants, and the clueless management types!
The human cloud works for sales people, and anyone involved with talking to other businesses, such as consultants and rate earners etc.
The people it doesn’t effect are all the internal business services such as accounts, HR, property management, production etc. In the grand scheme of things, they need to be at their desks – not at home or on the road etc.
What I imagine is going to happen over the next few years is that the internal IT is going to get outsourced so that whole companies are going to be private / public cloud based. The internal server is going to be housed in a data centre to keep it safe and updated. From a finance point of view, this could work out cheaper using jotted figures.
However, the unintended consistence is that now the entire internal private data flow is having to go from one desk, all the way along the public network to the data centre, and then all the way back to the private network and to the desk next door!
No, it isn’t the way to do it – but when there is a sale to be made, and a boss to impress, then its going to be happening all over the corporate world!
Eventually, we will return to our senses, and cloud will only get used where it is appropriate – just like we did when we realised not everyone needed a mobile phone!
Thanks, Oli. I’ve certainly seen a lot of business edge towards the cloud and as you highlight this trend will only increase over the next five years as the cloud’s infrastructure becomes more stable and the costs reduce. For many CEOs the financial savings here are too tantalising to resist. Indeed, one need only look at how the US government is using Amazon’s cloud solutions to see a successful example.
Yet it is this question of the public and private networks that causes many IT managers concern. Even though private networks are not 100% secure they do offer greater control and operational insight than cloud-based offerings. The question is which is more appropriate for your business and what data can you afford to live without?
It will be interesting to see how cloud technologies evolve over the next ten years. As you highlighted, technology moves in a very cyclical nature. In some respects, it’s only the name that changes.
How do you believe the cloud should be used? Should companies put more than just their CRM systems on it?
Thanks,
Gareth
New comment from Neil Davies on the Broadband Traffic Management LinkedIn group: http://linkd.in/iemwH7.
amusing article – we’ve been using ‘standard’ ADSL (256k or 448k up with 1mb or more down) to do all of this for the last 4 years – the bandwidth requirements are not large what is needed is delivering a consistent ‘experience’ to the data so that the user experience is acceptable. We are a completely ‘virtual’ company, all our connections are over retail network connections.
it is a matter of bit of performance/traffic engineering and appropriate deployment not more capacity. Our experience says that home workers (even ones with 30% of the companies distributed file-store) don’t mean a lot of data volume. Running such a remote office for a month of those consumes less data volume than watching a DVD once a week.
Thanks, Neil. I agree that consistency and a good quality of service are critical in the development of a mobile workforce. It’s great that you’ve been able to achieve this using today’s network connections. How data intensive are the applications you use? For example, do you use video conferencing at all? Earlier this week I was responsible for delivering a video conference with six offices spread across the UK, US, Germany, Australia, Singapore and Russia. The experience was less than ideal with frequent loss of audio and video signal.
Other problems also occur when trying to access CRM systems remotely, especially for team members who live in rural locations. The digital divide between urban and rural environments is continuing to expand and this also hampers the development of a remote workforce. In these situations, it is a case of connectivity and capacity that is required.
I was interested in your thoughts on traffic engineering. Are you suggesting that traffic management is the key to greater success here?
Thanks,
Gareth
New reply from Neil Davies on the Broadband Traffic Management LinkedIn group: http://linkd.in/iemwH7.
They can be very intensive (for example backups) so that the links are running, in both directions, continuously at 99%+ of capacity for several days.
We all have video phones on our desks – we use them several hours a day (e.g. complete presence while working on shared documents in our shared file store or with webex) – we have people in the UK and US – we met (and deliver consistently independent of the other traffic – ie when network is running at 99%+) our design target of better than three video glitch per active hour. All our voice system is VPNed and we only have VoIP based telephony at all sites for business. We also run 3G femto cells and our domestic requirements over same links.
CRM – we deliver a consistent service for that sort of thing as well, actually we’ve gone one step further so some apps (accounts) are only available in the cloud via RDP (to Window’s instances in AWS EC2)
I am particularly rural – 5.6km to exchange, DSL sync speeds that vary by 50% by time of day(or more phase of sun) – as low as 1200k occasionally as high as 2800k – but our approach adapts and manages the changes.
As you can most probably tell by this point I’m going to say traffic engineering is needed. There is a process you need to go through – but only once.
The good news is that you (in general) only need to do it at the edges (i.e. at CPE and PoP to access network) and you don’t require any active collaboration of the wholesale carrier.
Our setup manages itself – actually today is its 1000th day ‘anniversary’ of continuous un-interupted fault-free running for the PoP end , for us it is part of the furniture.
Thanks, Neil. It’s great to see your operations running so well. Congratulations on the anniversary. You appear to be a very forward-looking company and it’s encouraging to see such a virtualised organisation.
Have you encountered any difficulties along the way, particularly in regards to your WebEx or VoIP solutions?
Thanks,
Gareth
New reply from Neil Davies on the Broadband Traffic Management LinkedIn group: http://linkd.in/iemwH7.
The main issue is ‘isolation’ – you need to engineer the fact that (for example) the bursty traffic for WebEx doesn’t adversely effect the delivered delay characteristics to the VoIP, and that WebEx (for example) doesn’t destroy the normal operational stuff (other web browsing etc) – simple priority or bandwidth allocation doesn’t work well enough, you need more sophisticated queueing algorithms to get both the efficiency and the quality assurance.
New reply from Peter Thompson on the Broadband Traffic Management LinkedIn group: http://linkd.in/iemwH7.
I think I detect some incredulity in Gareth’s responses here, so let me say that I have seen Neil’s system in action and it does indeed work as advertised! The key point is that ‘standard’ queuing disciplines are unable to maintain appropriate QoS for multiple real-time applications such as VoIP and video conferencing when the link is fully utilised, so the expectation will be that you must ‘encounter difficulties’. The fact that you don’t is proof that a well-designed queuing/scheduling system can support reliable services without needing massive amounts of bandwidth.
Thanks, Peter. Apologies if you detect any sense of incredulity in my answers. This was not my intention. I’m genuinely impressed by Neil’s accomplishments and would be fascinated to learn more about Neil’s company and infrastructure. I still contend that global networks still have some way to go before a true mobile workforce can be realised. For example, yesterday evening I was visiting a small Yorkshire village and was scheduled to have a video conference with a colleague in the US. However, I had to drive over 30 minutes to find a stable network connection. Even then the connection was only capable of supporting voice and not video. How could people in this area be expected to work remotely?
Thanks,
Gareth
New reply from Neil Davies on the Broadband Traffic Management LinkedIn group: http://linkd.in/iemwH7.
Gareth,
The implicit point I was making is that it is not about the link capacity (once it is greater than about 256k up and 512k down – which is better that 98% of UK ADSL lines I read in an OfCom report somewhere) – it is *all* about how it ‘managed’ – this is an issue that ISPs/ Corporates/ End Users have complete control over, they don’t need active participation of the network wholesaler.
For all of our remote workers, be they 256k/512k (used to have one of those), or 400k/1000k (got one of those), or 832k/7M or even higher – without the traffic engineering there are unmanaged risks to service delivery that will bite you (and always at the worst time), with the right traffic engineering they all have assured quality of experience for their key services.
The difference between the sites is not whether services ‘work’, it is how many concurrent services they can run.
Neil
(currently 448k/2432k – worst/best downstream sync in past 4 months: 2048k/2880k)
New reply from Peter Thompson on the Broadband Traffic Management LinkedIn group: http://linkd.in/iemwH7.
Gareth,
Of course you need a reasonably stable network connection; you need enough bandwidth to meet the (inelastic) needs of the applications you need to run; and you need reasonable bounds on the intrinsic loss and delay characteristics in the absence of contention (satellite broadband might be a problem for VoIP, for example, because of the intrinsic speed-of-light delay). Neil’s example shows, however, that properly managed ADSL is good enough for voice, video and filesharing for a fully distributed small business, so the assumption that ‘work’ applications need higher bandwidths than are currently available doesn’t seem to be true. What they do need is smarter management of the bandwidth that’s already there.
Peter
Thanks, Neil, Peter. I appreciate your thoughts on this topic. I believe that we may be talking at slightly cross-purposes. When I discuss the development of the human cloud, I envisage a truly mobile workforce that can access online systems, such as communication tools or CRM platforms from home offices, hotels, airports, trains, etc. A workforce that can work at any time and from any location.
I work for a company with a global workforce of over 1,000 people. Many members of the team have home offices but spend 80% of their time travelling. In these situations they are continually struggling to find effective broadband access. I agree with you both that if you work from a static home office, it is possible to use traffic management to optimise your networking environment, but what happens when you’re away from that environment. This is where more robust and scalable access is required.
Thanks,
Gareth
New reply from Neil Davies on the Broadband Traffic Management LinkedIn group: http://linkd.in/iemwH7.
Ah,
Well the journey of 1000 mile starts with a single step….
To reach that type of Nirvana? at least 10 years, more likely 20 to 30… why?
The technology is all available it is about the lack of innovation in large corporations, the power of the incumbent, billing and provisioning systems (one of the largest barriers), cannibalisation of revenue fears, national regulation etc
Now if there was someone out there wanting to create a global, over-the-top, business critical class service to meet that need – it might shorten the timescales. I can’t see any way that the current players (and I know several of them intimately) could metamorphosise themselves into that sort of organisation – their corporate DNA just would not allow it.
Neil
New reply from Peter Thompson on the Broadband Traffic Management LinkedIn group: http://linkd.in/iemwH7.
So, to summarise: remote working using the existing fixed broadband infrastructure can work perfectly well providing the contention for the access link is managed using queuing algorithms that are fit for purpose. The pressure of ever-rising petrol prices may even break through the organisational/cultural barriers that stop most companies making more use of this potential at present!
Mobile broadband is a different ball-game: in addition to all the same contention management issues that apply to fixed broadband, there are the additional problems that (a) it uses radio, which is inherently susceptible to noise, fading, multipath interference etc. etc., (b) it’s a shared medium, so the available bandwidth varies depending on the other users in the same cell, and (c) it allows mobility, so the characteristics of the connection can suddenly change due to handover. All this notwithstanding the economic and organisational constraints that Neil speaks of!
So, I agree that this isn’t going to get sorted out overnight; I also agree that it isn’t about ‘bandwidth’ per se, but more about effective resource management that supports stable and consistent connections. And more cell towers, of course, which always have to go ‘somewhere else’!
New comment from Bob Yu on the IT360 Conference & Expo LinkedIn group: http://linkd.in/fTfP36.
Can’t wait for the Consumer Electronics Show, one of the most exciting moment of the year
.
Thanks, Bob. I’m continually amazed at the innovation on display at the show. Although early indicators seem to suggest that this year’s event may be a little quieter than previously expected.
Is there anything that you’re particularly excited about? I’ll certainly be watching to see the latest tablet news.
Thanks,
Gareth
New reply from Bob Yu on the IT360 Conference & Expo LinkedIn group: http://linkd.in/fTfP36.
Same, tablets will be something really interesting. So many big manufacturers are jumping inside this market. I will be interested in the RIM Playbook, so far it is very promising but I can’t wait to see some reviews. Anyways all those goodies are awesome but I wonder if they will supply enough battery power to stay on for a solid 8 hours? I doubt it.
Bob