Loosen the noose

The lure of sweet-talking cloud service providers, who boldly advocate infrastructure outsourcing, is rousing directional change in the mindset of IT teams.They were once petrified of entrusting precious company data and processes to anyone considered a third party, but that tenacious, do-it-yourself approach to IT no longer cuts it for the business owner smitten with the prospect of reduced costs, freed-up capital and heightened customer responsiveness.Increasingly CIOs are being expected to act like business change managers, with executive level conversations very much revolving around performance and risk mitigation. They become wise to the idea that efficiency and agility, or "business tempo”, as Geni-i’s Strategy Manager Neil Osmond prefers to call it, are best achieved in partnership with a service provider who is capable of expanding the breadth of the different cloud delivery models."They’re trying to avoid having mismatched SLAs and non-integrated, non end-to-end delivery of service,” says Osmond. "It comes down to four things: performance, cost, security and control.”Cloud services, and particularly infrastructure services anchoring a cloud, directly impact the operational heartbeat of the company, its business relationships, marketing and production, says Robin Cockayne, General Manager of infrastructure and services provider Revera. "Business must be light on its feet, to grow, shrink, or move sideways very quickly, without penalties.”On one hand, established legacy infrastructure, bureaucracy and limited flexibility is proving a noose around the neck of the CIO screaming out to launch dynamically scalable and virtualised resources. On the other, for every CIO enthused about increased business agility, there’s likely another arm-crossing defiantly, adamant that his existing service levels and processes are more than adequately addressing the needs of the business."Computing service providers should be sought out when they are thought to perform some aspect of IT better than the prospective buyer thinks they could do themselves,” says Cockayne. "Specific functions are bought back as a service and customer scenarios might include outsourcing all or part of an IT operation, ISVs and system integrators looking for a dependable hosting partner, or software-as-a-service (SaaS) providers wanting a scalable delivery platform.”You’re so immatureApprehension about the maturity of cloud computing has led some IT managers to question if it’s wise to take the plunge now, through fear of making duff investments that will need revisiting in a few years’ time. EMC’s CTO Arron Patterson sees the five-year business plan as impractical given that data quantities are escalating faster than the infrastructure can innovate to satiate its storage needs.Come 2020, the information in the Digital Universe will grow by a factor of 44; the number of "files” in it to be managed will grow by a factor of 67, and storage capacity will grow by a factor of 30. Yet the staffing and investment to manage the Digital Universe will grow by a factor of 1.4, says the 2010 Digital Universe Study, conducted by IDC on behalf of EMC. This should make it clear to CIOs and business executives that much of the next 10 years of their careers will be spent dealing with the mismatch of these growth rates.The amount of digital information created and replicated in the world will grow to an almost inconceivable 35 trillion gigabytes as all major forms of media – voice, TV, radio, print – complete the journey from analogue to digital. At the same time, the influx of consumer technologies into the workplace will create stresses and strains on the organisations that must manage, store, protect, and dispose of all this electronic content.Sales force automation, field service automation and HR related apps (like expense reporting and job leads) are the obvious leaders when looking at which complete applications to move to the cloud, says James Ketner, Product Design Director for AT&T’s Custom Cloud Solutions unit."What will be more interesting is what layers of the stack will end up in the cloud, or what workloads. For example, will the web server move to the cloud while back-end applications and database servers sit on dedicated hardware? Will the cloud be used to accelerate applications across a global WAN, or to offload a dedicated server that is running hot?”Decide which mission critical applications to shift to the cloud by addressing all of the interdependencies of those applications. They’ll be different for each usage case."There is the back-end IOPS that the application will need, as well as other traffic performance guarantees that are needed, from the amount of cycle bandwidth contention that could occur based on the cloud platform subscription ratio (i.e. Fibre channel bandwidth, Ethernet bandwidth and the Quality of Service (QoS) policies, or lack of), to the front-end user access models,” says Ketner.Buy-in at board level for technology investment of any kind is rarely straightforward. Cockayne asserts that early stage local adoption will be private rather than public, concentrating on niche on-demand services. "But once businesses have pooled computing resources, virtualised operating systems, floated applications and solved bottlenecks, then they can look at plugging extra capacity or services.” In any case, says Cockayne, business is moving that way because it is a more efficient way to run IT.We’re in the early days of maturity of cloud computing in general and service offerings are in the process of evolving. "People are still feeling their way,” says Osmond. "The benefits range anywhere from efficiency and speed to market, to better economics.”On-premise, public, private cloudsOn-premise IT is proving a popular route. Businesses own their own infrastructure and servers are located on their own premises. There’s also a market for those with data center housing, where businesses again own their own infrastructure, but house it at a service provider’s data center. The large grey area in the middle, says Gen-i’s Osmond, is defined as ‘pure public services’.Public cloud would be defined as Google, Amazon, Facebook–the very large-scale dedicated servers somewhere out there in the Ether that you can access via the internet. Between data center housing and public cloud, lies private cloud.Private cloud is dedicated to the enterprise, so it’s not shared. "It is basically a virtualised utility platform, or utility infrastructure, that is not necessarily paid for with CAPEX, but is more of a lease,” says Osmond. Larger public internet clouds are ideal for template applications, such as remote backup, but in whichever direction things go, clouds must protect and guarantee outcomes. "Practicalities such as data availability, sovereignty, prevailing law and New Zealand’s reliance on a single international internet gateway, will concentrate local cloud computing models to strata rather than all-embracing services over international internet,” says Cockayne.Yank those purse stringsAs the global economy has been shaken, small and medium sized businesses are the hardest impacted on their average profit margins, so they are looking at ways of increasing their return on assets. They perceive cloud as providing the best way to reach that higher value, says Ketner. "By converting their fixed cost to variable costs they have the opportunity to increase the ROA. Cloud is much more than outsourcing all or portions of your IT infrastructure to a third party cloud service provider, it is a business objective to gain higher returns.”When assessing which systems to move from a physical environment to a virtual one today, at low cost, Cockayne advises CIOs start by assessing the four main IT costs: environment, people, software and hardware. "You can’t really avoid software applications and you’re going to pay for environment management one way or another. But once you have all your IT in one place and tap virtualisation and pooled capacity, efficiencies are there for the taking, including a lower ratio of IT workers to machines, improved server and storage capacity utilisation and streamlined backup.”Once a vision and sponsorship around using cloud services has been established, together with a productive governance framework around how the business will use these services, a CIO should look at the ongoing funding model. "They would probably look at moving that from being project based to being more of a pooled funding model, because you’re moving from CAPEX to OPEX,” recommends Osmond."The advantage of on-demand virtualised provisioning is in service management peripheries: networking, firewalls and automated backup. All the things on the perimeter are relatively cheap,” says Cockayne, "because your capacity is drawn from a consolidated platform. Do-it-yourself systems are capital intensive and come saddled with an operational overhead. It starts off looking cheaper but ends up costing more longterm”.The much-hyped cost model is pay-as-you-go, says Gen-i’s Osmond, but that is frightening some business operators who prefer the surety of billing. "They actually want to have a 36 month agreement because it’s less likely to invoke bill shock.” Pay-per-use on a month-by-month basis is not as popular as perceived, says Osmond. There are a few who prefer not to become bound to a five-year lease, but equally, a fear of commitment to a fixed term agreement is not commonplace. Businesses are keen on set terms, but are looking to timeframes of six to 12 months, rather than 36-month agreements.Will I lose my job?A process of transformation follows, because an IT environment centered around ‘build and operate’ and ‘deploy and purchase’ components needs to be transformed in how it operates with providers in a utility sense. "You’re probably going to have less technical code writers and more relationship contract management people,” says Osmond.Cloud service providers build their business case on extracting the complexity away from the end user. All the user needs to worry about is the interface, so the PC screen or the handset, for example. The technicalities of supply should no longer concern IT managers – that’s the domain of the supplier, creating a huge opportunity for system integrators.It’s this element of cloud computing that injects fear into the hearts of IT teams. With repetitive IT administration tasks automated, what happens to the techie as we know him/her?We’ll find new skillsets emerging in the enterprise, says Osmond. They’ll be remodelled to adapt to the process change. The Platform as a Service concept means the task of infrastructure management is handed over, freeing up time to use programming skills to rapidly write applications. There’ll be less need for a six-month application build process – timeframes will turn into weeks, or days.Every 18 months the process and capabilities of cloud computing are likely to double, says Osmond, so without doubt, in three years’ time, a business will get better efficiencies out of virtual infrastructure than we do today, but "the utility concept is that I, as a customer, let someone else worry about the lifecycle upgrade, streamware and replacements.” The service provider owns the risk around the infrastructure and has to build replacements into the price."Engaging enterprise data center providers should be with a view to shared services, or infrastructure-as-a-service (IaaS), which is where the big savings are,” says Cockayne. "Co-location is a legitimate transitional strategy, but by no means the end game. Co-location on its own is unsustainable and the sooner organisations recognise its shortcomings the better off they’ll be."Co-location’s dubious sustainability is in some cases overshadowed by unwitting providers who invite disaster by relocating client systems to "high-risk event zones”, including the Wellington waterfront and geologically suspect sites. Unsustainable? Yes. Irresponsible? Most definitely.”Resellers will need to adjust their positioning to sell annuity-based solution rather than box-dropping, followed by "tradesperson” support models, says Cockayne. "Centralisation will affect their market considerably in this decade.”Keep it in the familyTo store onshore or offshore? That is the question being asked by businesses most concerned with the accessibility and security of their data. Data sovereignty has gleaned some attention in recent weeks, after the Inland Revenue issued what’s been branded a sensationalised caution about the storage of business records. Only business records stored in data centers physically located in New Zealand comply with the record keeping obligations of the Inland Revenue Acts.Section 22 of the Tax Administration Act 1994 says that any business in New Zealand has to keep sufficient records in New Zealand (and in English) to enable the commissioner to readily ascertain information about their tax affairs.Twitter lit up with many users asking what this meant for cloud companies such as Xero.Rod Drury, CEO of Xero, was quick to respond to the Revenue Alert, posting a blog that put out any flames: "A few tweets just came in noting that the IRD has issued an alert concerning location of data for cloud computing,” Drury said. "There is no issue for Xero customers as we have been working with the IRD for some time on a blanket dispensation and assisting IRD on policies around the cloud.”Drury wrote that New Zealand legislation hasn’t kept up with developments in technology compared to other countries: "We are working towards certification of our current customers and in the longer term expect to see the legislation amended to better reflect contemporary technology. We’d expect this to end up in a similar position to Australia where there is no onshore storage requirement, only that your records are available if requested.""Corporations and organisations should be held accountable for their own data,” says Ketner. "I think an organisation will have to enhance their information management to determine what is "cloud” and what is "local”. We are a global economy and data will just become another import /export regulation.”Revera’s Cockayne presents a fiercer argument for keeping data within New Zealand. Gen-i’s Osmond shirks the idea that data sovereignty is actually earning much talk time amongst his clients. Hosting overseas is "a bit like someone promoting a flash new data center in Wellington, but locating it on a fault line. The so-called extra security is undone by greater risk factors. Why go there?” asks Cockayne."If you look at value in and out, using overseas companies out of New Zealand should be about as popular as a bank executive bonus programme."Yes, they employ Kiwis and feed the local services pool, but the spotlight of overseas head office accounting turns attention to profit and loss ploys, making little or no profit to minimise local taxes and transferring losses to limit overseas tax commitments. There’s a good reason why companies like IBM are relocating information and processing to the homeland –control. Information and the processes that give it life are the new oil fields. Choose wisely. A few cents saved on processing is nothing compared to the pain of helplessness."The bottom line is only go overseas when you can’t get what you need, at the right price, in New Zealand. There is no reason why local cloud services should cost more than their overseas equivalents,” says Cockayne.Essentially, the decision comes down to risk. Hosting data offshore requires a comprehensive analysis of compliance, security and control.