The channel impact of cloud computing

Will the domino effect undermine traditional business?

(Thoughts for Distributors, Resellers and System Integrators)

The writing is on the wall. As cloud computing steamrollers across the market, on-premise infrastructure will be the first to go. And with nothing to implement or look after, there’ll no longer be a need for in-house IT staff. Then, of course, the traditional channel business will collapse, as there will be no one left to buy anything. Oh yes, and after all of the dominos have tumbled, Amazon, Google and the like will inherit the earth.

Well that’s the line we hear from cloud evangelists and suppliers trying to drum up business for hosted services of one kind or another. If you subscribe to this view, then you’d better stop reading now, as we are not going to entertain such nonsense in the remainder of this article. Instead, we are going to apply some common sense and take a more grounded look at how cloud computing in its various forms will co-exist with more traditional approaches to solving business computing needs.

Given the ambiguity around the meaning of cloud, for the purposes of our discussion we are going to consider two categories of offering. Firstly, ‘private cloud’, which allows organisations to aggregate server and/or storage equipment in their data centre or computer room to form pools of virtualised resources. Compute power or disk space can then potentially be allocated flexibly on-demand, boosting overall efficiency and enabling IT to respond more quickly to new requirements and change requests from the business.

Private cloud is an evolution of the activity we have seen around virtualisation. At the moment, it is the form of cloud computing in which mainstream customers are most interested. With equipment and software running on-premise, the opportunity for resellers and SIs is clear, especially those targeting mid-sized and large organisations. As we are still in the early market, however, exploiting the opportunity will require more of a consultative/educational approach. Fortunately, key vendors are now well geared up to help channel partners with cross-sell and up-sell techniques and collateral.

The second category of cloud we shall consider is on-demand hosted services. Within this, we find a range of offerings such as Infrastructure as a Service (IaaS), Platform as a Service (PaaS) and Software as a Service (SaaS). If you want an explanation of these, and other forms of hosted service that fall under the cloud banner, then check out our recent paper ‘Applied Cloud Computing’. Suffice it to say for now that similar technology advances to those underpinning private cloud are allowing service providers to deliver more flexible and cost effective hosted services of all kinds. This, in turn, is shifting the line in terms of what is feasible or attractive for customers to outsource to a third party hosting company.

Despite the possibilities, the demand for hosted cloud services, whilst growing, is still relatively small compared to the overall activity in the IT marketplace. Furthermore, rightly or wrongly, a general lack of trust in the model continues to act a brake on mainstream progress. With a deep cultural dimension to this, the situation is not going to suddenly change. We are therefore unlikely to see a wholesale shift in emphasis from on-premise to hosted solutions overnight.

Even as the comfort level grows, the market can only shift so far. For all but the smallest customers, the notional ‘move to the cloud’ would typically involve subscribing to many services from many different providers to meet the wide and varied needs of the business. The integration, maintenance, support, accountability, and supplier management issues that arise when dependencies cross many different operational domains are daunting. So too are the practicalities of coordinating security and compliance policy across jurisdictions.

Things will get less scary over time as interoperability and service assurance standards emerge, and as the service provider community matures beyond the current ‘every provider for themselves’ mentality that prevails. The continued evolution of monitoring and management tools will also help to keep things working as smoothly as possible across a distributed on-premise/provider landscape. All of this will simply reduce rather than eradicate the risks, however – hence there being a natural limit to how much of a shift to cloud services will take place in most organisations.

Having said this, as we discuss in the aforementioned paper, cloud computing is creating both choice and complexity for customers, so it cannot be ignored by those in the IT channel. Skilling up will therefore be necessary to maintain account control and exploit the spin-off business. There may also be some incremental margin to be gained from hosted service resale or referral. For small customers in particular, it may be possible to meet needs through hosted services in areas such as messaging, collaboration and CRM that would otherwise not be feasible for commercial or practical reasons through the on-premise alternative.

The bottom line is that there is no need for resellers and SIs to panic about the rug suddenly being pulled from under them by the emergence of cloud. However, doing nothing about it could represent an increasingly high opportunity cost over time, especially as more service providers figure out the importance of the channel in reaching and supporting customers. There are therefore definitely advantages to embracing cloud computing proactively.

Taking advantage of the Windows 7 migration

After almost a decade of service, the end-of-days is drawing near for the long-lived and very widely used Windows XP. It has entered legacy mode with the end of mainstream support. The extended life of Windows XP was partly due to economic conditions hampering spend on desktop upgrades and partly down to Windows Vista failing to get off the ground successfully in the business sector; it never really recovered the confidence that was lost in the early days because of compatibility, performance and usability issues.

Nevertheless, we are where we are, with many organisations relying on a desktop operating system designed over a decade ago that is facing obsolescence, with all of the direct and third party software support and maintenance implications that come with that. It is a position that is not sustainable for much longer, unless you are willing to live with escalating cost and risk [link: http://www.freeformdynamics.com/fullarticle.asp?aid=848 ]. Every day extra spent running Windows XP adds additional risk to the business as support costs increase, security issues escalate and software vendors target their products and support at newer Windows releases.

With this in mind, while other options exist, the most obvious way forward that most organisations are considering is a move to Windows 7. This has been pretty well received and represents a massive turnaround. It is Microsoft’s fastest selling operating system and has become widely adopted by consumers and also small businesses in the year since its release. Now we are seeing signs that mid and large sized companies are preparing to make the move too.

Windows 7 is a derivative of Windows Vista, albeit a much improved one, so benefits from a lot of the ground work done in terms of compatibility, interoperability, security and manageability. While it has only been around for a year, much of the componentry and core feature set has been road-tested and optimised for significantly longer than that.

So is it just a case of out with the old and in with the new? Well clearly not. Despite the general level of comfort with Windows 7, there are many practicalities to still consider.

Doing a straight switch from Windows XP onto Windows 7 may seem like a simple approach, but migration in practice takes extensive planning, testing and remedial work, which is costly and time consuming. This work needs to be performed on legacy applications as well as new ones, as well as looking at the underlying hardware. Beyond this familiar territory, if you think a little bit ahead and consider also other changes that are happening to the PC, it may help to alleviate or avoid issues that if retrofitted at a later date could increase risk and costs and decrease the capabilities of the solution.

On the software front, the Windows 7 migration is a good opportunity to evaluate, test and introduce up-to-date versions of software. This is true for specialised software as well as more general productivity suites, and particularly Office. Many of the companies that we have surveyed are still running Office 2003. Although it is still reasonably functional, it is showing its age with a lack of support for new document features and standards, less capability to support mobility and a higher incidence of security issues.

Like Windows Vista, many companies avoided the big change that was Office 2007. Office 2010 continues the evolution of Office, and enhances much of the interaction with back office functions such as Exchange, SQL Server and SharePoint. After the upgrade licence costs, testing and user training are some of the biggest expenses involved in upgrading Office. As extensive training and testing have to be done as part of the Windows 7 migration, it makes sense to take advantage of the activity to support an Office switch.

Looking at the underlying hardware, there are a couple of big changes to consider. The first is whether to move to 64-bit support or not. Windows 7 has a 64-bit version that can perform much better with increasing memory capacities. The beauty of it is that, from an end-user perspective, it is almost indistinguishable from the 32-bit flavour. So the issue is – is it best to jump straight to 64-bit at the same time as adopting Windows 7, or to wait and only migrate when applications demand it?

The advice here is to leave your options open. As applications and hardware may behave differently, it is definitely worth testing them on both 32-bit and 64-bit for compatibility. This will identify any issues that may cause problems when moving, and allow workarounds to be considered – for example using the built in Windows XP mode of the business versions of Windows 7.

Ultimately, most applications are likely to lag the operating system by a number of years in moving to 64-bit. This may seem like a compelling reason to stick with 32-bit for the time being. However, without moving the platform to 64-bit there will be no opportunity to run applications that have been updated to take advantage of the extensions without a time consuming and costly re-installation of the operating system.

Beyond performance and compatibility, the ever-increasing move to mobile working and the more sinister nature of threats and malware means security is high on the agenda. An upgrade presents the opportunity to introduce additional security in the shape of encryption for data and information. The encryption functionality can add significant overhead, so the introduction needs to be carefully managed. It may need to be implemented in conjunction with hardware upgrades – particularly modern CPUs with encryption acceleration or Solid State Disks (SSDs) - to maintain performance.

Whatever your migration plans for Windows 7, it is important to look at the wider view to take advantage of the biggest transition in years to improve and optimise the capabilities of the PC pool as retrofitting at a later date may be expensive and time consuming to implement.

Originally published on CIO (registration required).

Taking it to the n’th degree

Why you should upgrade to 802.11n

The notebook market started to gather steam in the early 2000’s with the launch of mobile initiatives such as Intel Centrino in 2003. This spurred the development of an installed base of Wi-Fi clients where previously uptake had been low. Although some companies took a proactive approach to building out Wi-Fi networks, many companies either reacted to demand with a piecemeal infrastructure or were surprised to find that they had many unauthorised access points installed by ‘enterprising’ users.

The performance of Wi-Fi, particularly in the early days, could be characterised as sluggish at best, and the wired network was preferred in most instances where possible. With the Wi-Fi network as a convenience, the approach to integration, management and security was often treated as an afterthought.

The introduction of the second generation of Wi-Fi, 802.11g, was good in that it lifted performance a notch, but this in itself was not compelling enough to encourage proper joined up investment in infrastructure. The pressure to upgrade was tempered by the presence of legacy devices that could not run faster, or would interfere and result in general slowdowns on faster networks.

In the meantime, network connectivity has become so ubiquitous that it is now a critical element that many applications need in order to achieve full functionality. On the PC front, user expectations of performance have escalated as gigabit networks have become widespread, while the big swing to notebooks means that many employees are using them away from the desk and default to wireless.

On the mobile side there has been an explosion of new devices such as Smartphones and Tablets that rely entirely on wireless for connectivity. These devices are capable of running sophisticated applications and consuming rich media, just as the traditional PC is. This need for Wi-Fi as an equal partner to wired connectivity is driving the development of the blended wireless edge network.

All of this, together with the escalating costs and risks that trouble existing installations, brings the 802.11n, the latest and greatest version of Wi-Fi, to the top of mind. Unlike its predecessors, this version is both fit for purpose - with a throughput comparable to Fast Ethernet and increased coverage to support more devices with fewer access points - as well as future proof as it has only recently been ratified and no successor is immediately around the corner to wait for before splashing out.

Another aspect to consider is that although the standard has only recently been ratified, devices based on the draft specification have been in the market for years. This has resulted in a large pool of 802.11n client devices that can benefit immediately from an updated infrastructure. At the same time, many of the devices with legacy Wi-Fi (802.11b and g) are approaching the end of their life and being removed from the active client base.

With up to date hardware and software that is being actively developed and supported, 802.11n is also easier to integrate from a security and management perspective. Older Wi-Fi kit suffers from security that is less than robust, with WEP in particular being easy to crack. With modern Wi-Fi infrastructure, access and encryption are far more secure with WPA2, especially when 802.1X is used for authentication.

Now is the time, if your Wi-Fi infrastructure is based mainly on older standards, to consider updating to 802.11n. Upgrading the Wi-Fi network on its own will bring many advantages, but there is also the opportunity to reconsider how the wireless and wired networks should work together, with increasing integration and compatibility.

Whether you choose to integrate the two or operate Wi-Fi as a standalone network is up to you. Integrating the management and policy across the two may not be a trivial matter to undertake, especially in multi-vendor networks. However, the demands on the wireless network may mean that developing separate management or security tools and policies, such as Network Admission Control (NAC), for each network would be cost-prohibitive. The initial pain and cost of integration may well be worth it for the longer term benefits. In some cases, the wireless network can help to develop the wired one, as in the case of 802.1X which is in widespread use on Wi-Fi networks but has yet to gather momentum on wired networks.

Whichever path you choose, the time is right to do something for all of the above reasons. If you do it right, you’ll not only be ahead of user demand for once, but will also position yourself well for the future.

Originally published on Computing.