WAN Governance

7 significant digits (?)

The pricing of Barcelona Cathedral's public park is of high precision… For the last digit to represent a useful value (say 10 cents) a car driver should park her car for (approximately) 100,000 minutes, i.e. 69 days… but then the daily price would have applied, wouldn't it? A mere 5 cents/minute should be more convenient and as protective of car owners' finances.

When looking to the average business case of WAN Governance, we find similar situations where using round numbers and order of magnitude are easier to use and finally more significant.

Don't you think that "For $3/employee/month you save $30/employee/month and guarantee the performance of an application portfolio that cost enterprise $300/employee/month" is an easier and more useful guideline than "For $2.72 to $4.25/employee/month you save $24.1 to 32.8/employee/month and guarantee the performance of an application portfolio that cost $142 to $433/employee/month"?

Naturally, when evaluating the WAN Governance business case to particular enterprises, we use their actual data, organization and priorities to make "sur mesure" studies and evaluations. But even then, don't expect us to provide figures with 7 significant digits…

Béatrice Piquer-Durand

In a similar vein, you might remember the hilarious Seth Godin's talk at Ted?

Unified Communications – A perfect WAN Governance Business Case

Unified Communications (UC) offer huge opportunities for the enterprise, with benefits including improved productivity, easier collaboration, better customer satisfaction and ultimately, higher profits through more efficient teamwork and competitive advantage. But UC come with their distinct challenges and not every project yields the anticipated ROI.

Challenge 1: UC implementations are a significant investment. If they are to yield a good ROI and deliver a suitable user experience, then their performance must be guaranteed as they rely heavily on the Wide Area Network (WAN). Recent research about Unified Communications deployments demonstrated an average Total Cost of Ownership (TCO) of $36 per user and per month, without WAN circuit costs and large distance charges.

Challenge 2: UC are made up of data, video and voice flows with dynamic, variable traffic patterns that grow or subside based on user demand. A single desktop video flow means upwards of 300-500 Kbps between branch office sites. In turn, this can result in UC applications impacting the performance of other business critical applications during peaks of high demand.

Challenge 3: UC encourage desktop to desktop video communication which is wonderful for productivity, but they place enormous strain on inter-branch WAN links that carry this traffic. Traditional static Classes of Services (CoS) are not able to match such complex traffic situations and overall business application performance is put at risk.

Scenario – Let's consider a company with 5,000 employees in the Professional Services industry that is deploying Unified Communications. The company has an IT budget aligned to its industry standard, i.e. $975/user/month (among which $320 is allocated to applications). A full featured WAN Governance system will typically have a TCO of $3/user/month (including central devices, management opex and unforeseen project incidents), for the following benefits:

• Guaranteeing the quality of the UC applications (an investment of $36/user/month);
• Protecting the performance of the other business critical applications (ERP, CRM, file sharing, cloud apps, Internet, etc.) against resource intensive UC (an average investment of $320/user/month);
• Automatic incident prevention, better helpdesk productivity and IT governance will save a very conservative $15/user/month;
• Delaying bandwidth upgrades by a few years, saving an average of $15/user/month;
• Optionally deploying hybrid networks for accessing large resources at low cost (e.g. desktop video), dividing the cost of incremental Mbps by 3 to 5.

Not deploying WAN Governance would jeopardize not only the success of UC deployment but also the performance of the other business applications. Unsatisfied users and business managers will complain, lowering UC adoption and forcing IT to upgrade the WAN at a high cost (without being sure of the impact).

Alignment with the company's strategic goals – Guaranteeing the success of its UC project, WAN Governance leverages the company's investment to improve its efficiency and productivity through better collaboration. Moreover, enterprises in the B-to-C sector will significantly increase their customer satisfaction and their competitive advantage.

Béatrice Piquer-Durand

The Economic Case for Cloud Computing (Part 2/2)

(…/…)

Hidden Costs of Cloud Computing - So much of what makes the cloud beneficial to companies – its scalability, its flexibility, its collaborative nature – also adds expense. The fact that IT infrastructure is not in-house means that it's owned by someone other than the company, creating compliance and security risks. The storing of data on foreign servers creates concerns about compliance with privacy standards and regulation of data access. The increased use of cloud-hosted applications creates stressed networks. All these things, by requiring additional resources and manpower, add expense to cloud computing.

Securing Data in the Cloud - Security is an undeniable concern for companies. R&D procedures, clinical trials results, marketing information, industry sales details, and other company-sensitive information must be protected from falling into the wrong hands. Owing to the external nature of cloud computing, where third-party users may be managing servers, unknown personnel may have difficult-to-monitor administrative access privileges. Companies offering cloud computing don't simply allow these security concerns to go unaddressed; most provide extensive firewalls, security codes, and other business practices to build customer confidence and guarantee protection. Yet companies must likewise be active, through encrypting data, monitoring the activities of privileged third parties and demanding greater visibility from cloud providers about how they are securing hosted information.

Compliance of Data Hosted in the Cloud - Most data compliance laws and regulations were written under the assumption that the liable party controls both the infrastructure where data is stored, and the ability to make decisions about where that storage is located. This means that most compliance laws assign responsibility to the user of a cloud computing environment. Such laws have, in the past, tended to limit activity in the cloud to the manipulation of data that doesn't involve intellectual property or patents. To ensure compliance, companies must monitor the infrastructure of the cloud, made difficult by the ever-shifting, adjusting nature of the cloud. As with security, this activity requires a monitoring and service assurance operation to be established.

As IT Becomes less Complex, Networking Complexity and Cost can Increase - By moving to the cloud, companies simplify and streamline IT, concentrating it in a few locations – some of which they control, some of which they don't. This means that the link between the user of the data, and where the data sits, is more complicated than before. In a nutshell, the network is more important and vastly more complex than ever before. If companies can't deliver data, the economics of the cloud will be reduced. A wealth of research and information hosted in the cloud will provide no benefit if it can't move to where it needs to be. Likewise, applications hosted in the cloud will be pointless if a CIO can't guarantee their performance. If a company has 30,000 employees working from other offices, it's important they aren't waiting behind their laptops for applications to work.

According to Gartner, average IT costs represent €800 per employee - and the cost of the networks that support the delivery of that IT is only a fraction of the total. Nevertheless, a move to the cloud necessitates investment in the network. As users access more cloud applications, networks can turn into the London M25, with too much traffic competing for network resources. In this situation, rather than adding new costly lanes to the motorway, savvy firms are protecting the performance of business critical applications by practicing WAN governance. The basic premise is to build greater intelligence into the network so that when network resources are assigned priority is given to those users of business critical applications such as an ERP system, rather than bandwidth hungry applications like YouTube or Facebook. You can think of this like adding a priority lane to the motorway, or how Transport for London is planning to help athletes move around the city during the Olympic Games next year.

Conclusion - It is inevitable that business is using the cloud in some capacity. Users accessing applications such as LinkedIn or YouTube are already accessing this software as a service. Gartner, the noted IT analysis firm, has predicted that cloud computing revenues will top $150 billion by 2013 as a large number of firms change their IT model. When considering the cloud, it's important to remember there is no IT nirvana and not all costs are immediately obvious. Cloud computing requires a change in mindset as technical teams become responsible for guaranteeing service and application delivery, rather than 'doing' the nuts and bolts of IT.

(A full version of this article has been previously published on Samedan)

Illustration: « La fée électricité » (details), by Raoul Dufy (1937) – It is a really big painting… and other "fairies".

The Economic Case for Cloud Computing (Part 1/2)

In Nicholas Carr's work The Big Switch, the development of IT is likened to the development of the modern electric utility grid 100 years ago, when companies stopped generating their own power with steam engines and dynamos and plugged into the newly built electric grid (1). Today, a similar revolution is underway. Hooked up to the internet's global computing grid, massive information-processing plants have begun pumping data and software code into businesses. According to Carr, computing is now turning into a utility. But how will cloud computing change the economics of IT and is it really as simple as signing up to low cost computing-as-a-service?

Reduced IT Capital Expenditure - This is one of the most cited and most immediately obvious benefits of cloud computing. By allowing individuals to access centralized IT resources using network-based computing, cloud reduces the need for investment in hardware and/or software at business locations. Combined with outsourcing of infrastructure and adoption of software-as-a-service (SAAS), up-front hardware and software license costs can be exchanged for a rental model, like any other utility service. The something-as-a-service (XAAS) model means that infrastructure, software and platforms are all provided by third party IT suppliers, effectively removing the need for any up front IT investment.

IT Economies of Scale and Efficiency - Synonymous with cloud computing is centralization of the IT estate and servers, either in a company's own data center or that of a cloud provider. In this circumstance there is the ability to virtualize the IT environment, and this means a server can run multiple applications rather than just one, as was traditionally the case. For instance, say a pharma company purchases a server to utilize in traditional IT models. Regardless of whether or not this server is being utilized, the ongoing cost still exists, with the machine often running well below capacity. With virtualization in a centralized environment, other jobs can be shifted onto the spare capacity of dormant servers and it's possible to run servers at over 90 per cent utilization.

The Opportunity to Contribute - The cloud enables communication at all levels: across the office, across the hall, across the nation and around the globe. Complex genetic sequences and biomarker data are being hosted in the cloud at an increasing rate. Data is available for access by pharma companies according to research needs. This allows an increased collaboration of existing knowledge – and the financial benefit of not having to conduct research when that research has already been done. For instance, advances in nanoscale and microfluidic chemistry means that DNA can be monitored by photographic sensors, producing collected images of up to 800 gigabytes. According to Chris Dagdigian, principal consultant for BioTeam Inc, "this creates a massive data-capture and handling problem. We are now in an era where instruments that are showing up in very small wet laboratories are capable of producing a terabyte or more of data in a day (2)."

Scalability - The ability to expand or contract servers as necessary allows companies to pay for computing time as they need it, rather than maintaining their own servers. With the cloud, using one server for 1,000 hours costs nearly no more than using 1,000 servers for one hour. Such flexibility is, in a sense, priceless. Take this example from Michael Naimoli, Microsoft's National Director of US Life Sciences, about programs to process protein folding (3): "Look at all the data that goes into protein folding. Companies are currently developing large molecule products – usually called monoclonal antibodies. The activity of that molecule is bound up in how it folds itself. During the discovery process they like to look at the primary sequence of that product and they want to do calculations about how it's going to fold. They have traditionally maintained a large number of CPUs that have to be spun up around that activity and it can take 70 hours to finish up the whole protein folding analysis.

Maximizing Efficiency of IT Staff - This is one of those 'hidden' economic benefits of cloud computing, something not immediately recognized, yet still very much a reality. Maximizing the efficiency of IT people occurs in two ways. Firstly, companies make better use of staff as a byproduct of concentrating servers in larger datacenters. This concentration means fewer workers are needed for infrastructure management. There are fewer people per server, freeing up staff for other activities and resulting in financial savings. Secondly, cloud computing allows a company to reduce the mix of skills it maintains on the payroll. Most companies do not want to acquire skills for the management of all applications they're likely to use – the skills to host and manage different applications, for instance, which can require lengthy training of staff. As hosting and managing exists in the cloud and is more often than not delivered by a third party specialist, companies are free to concentrate their energies on what matters: using the applications.

Minimizing Response Times - Using the cloud is fairly simple. This ease of use, combined with the ability to communicate with a team anywhere across the globe (as information and applications are not tied to a particular location), provides certain economic benefits. Companies can stop or start processes almost instantly. Such rapid responses can be the difference between funds gained or lost. Consider a quote from Eli Lilly's Dave Powers, the firm's associate information consultant for discovery IT, on the firm's usage of Amazon's Elastic Compute Cloud service: "We were recently able to launch a 64-machine cluster computer working on bioinformatics sequence information, complete the work, and shut it down in 20 minutes. It cost $6.40. To do that internally – to go from nothing to getting a 64-machine cluster installed and qualified – is a 12-week process (4)."

(…/…)

References

1. Carr N, The Big Switch: Rewiring the World, from Edison to Google, 2008
2. Mullin R, The New Computing Pioneers, Chemical and Engineering News 87(21): pp10-14, 2009
3. Koroneos G, Pharma and Cloud Computing: Are We There Yet?, http://blog.pharmexec.com/2010/06/16/pharma-and-cloud-computing-are-we-there-yet/, 2010
4. Mullin R, The New Computing Pioneers, Chemical and Engineering News 87(21): pp10-14, 2009

Illustration: « La fée électricité » (details), by Raoul Dufy (1937).

Consolidation: it’s the apps that matter

A post from Nigel Pink, VP and general manager UK, Ipanema Technologies

As a result of acquisitions multinational enterprises often have a plethora of different IT systems. Data centre and server consolidation simplifies complex IT infrastructure and reduces IT cost but increases complexity when delivering applications over the network.

The costs of distributed enterprise infrastructure include hardware, software, maintenance, security, and backup. These costs dramatically increase with the number of disparate systems. Global businesses operate branch offices and data centres at various domestic and international locations and, in addition, are increasingly employing mobile workers. To consolidate IT infrastructure, enterprises typically choose between one or two data centres (particularly for disaster recovery) to which existing applications as well as new applications are migrated. Even though these applications are now hosted in increasingly remote data centres they continue to be used in the same way as before. Practically, this means that enterprise information no longer flows locally in the LAN, but rather over the WAN resulting in slower response times and a considerable increase in data traffic. Ultimately, if unprepared, IT consolidation projects can unwittingly cause application performance to be degraded, impacting user productivity.

More bandwidth isn't the answer to user satisfaction
Bandwidth can be an expensive IT cost. Removing server, hardware and maintenance costs through consolidation loses much of its impact if you then need to double network bandwidth capability to support the distributed infrastructure. As we have found working with lots of organisations bandwidth can be consumed disproportionately by non-critical applications. At one oil company a single user was consuming over 90% of the bandwidth at a site through the use of recreational video applications. A WAN must have controls built in that can allocate resources based on the criticality of the application in question. For example, you must be able to specify objectives such as 'any end user session, anywhere, anytime, must have at least 50kbps if using SAP'. The Ipanema system works at this granular level and that's how we can help organisations to specify SLAs for each application at the individual user level. In the industry this is sometimes referred to as Application Performance Management (APM) but at Ipanema we prefer the term WAN governance.

Consolidation and the WAN
Consolidation of servers and applications means more data traffic flowing across the WAN. As a result application performance suffers due to delay, jitter, and packet losses. According to a study we commissioned in the UK, the cost of poor application performance can amount to additional costs upwards of £500,000 per year, due to reduced productivity, unfilled orders, and decreased customer satisfaction. The research also showed that network bottlenecks resulting from centralisation can generally be eliminated by increasing bandwidth. Although 57% of respondents confirmed this, 75% feel that increasing bandwidth alone does not solve the application usage problems that occur from increased centralisation of IT resources.

Significantly, 93% said that they wish for improved network visibility and control of the data traffic flowing over the WAN.
Access to centrally stored remote information, groupware systems such as Outlook, Sharepoint or Lotus Notes, and important proprietary applications like SAP are all vulnerable to performance issues post-consolidation. WAN Governance allows applications to operate according to their level of business criticality and can guarantee the performance of critical applications at the individual user level.

Governing WAN Infrastructure – it's more than acceleration
Current enterprise services run smoothly in the local environment, but were not designed for operation over a WAN. Using the WAN to access applications significantly increases bandwidth requirements between sites.

To solve this problem a variety of WAN optimisation tactics can be deployed including; compressing data and local caching at various locations, accelerating the transfer protocols used on the network, changing the network topology of information flows. However this tactical acceleration approach alone isn't enough. Imagine a Ferrari travelling round the M25 at top speed – it looks impressive. But now try it in rush hour and the Ferrari can't work to even a fraction of its potential. The same is true of an acceleration system deployed on a congested corporate WAN. Acceleration is part of the battle but it must be viewed within the wider context, which is where WAN Governance comes in.

The WAN governed as a whole thus becomes a solid asset on which the business can rely at all times. When conducting consolidation projects we urge the industry to consider the delivery of applications up front. After all users won't appreciate the IT efficiency savings you've achieved but they'll sure notice if their applications grind to a halt.

(This article has been previously published in Datacenter Management)

Illustration: The gluttonous Gargantua, as depicted by François Rabelais and etched by Gustave Doré.