 |
|
 |
 |
 |
“Our unique selling point is that we model the clients’ use of their infrastructure, networks and assets more rapidly, economically and easily than any other provider.”
Peter Meehan Director Business Development of AAL UK |
 |
 |
The language of the IT world includes many impressive words like rationalisation, consolidation and virtualisation, but they can all be summed up by another much simpler term – change.
Technology moves so quickly that it spawns constant change. Enterprises regularly rationalise their datacentres or hardware; consolidate or refresh their networks, alter platforms, build asset registers or move up to next generation applications or networks. They also need to constantly scrutinise their infrastructures to ensure optimum business efficiency for both performance and cost.
This is a good sign because it shows that enterprises are looking to the future – but it can be difficult to make the right decisions for tomorrow if you don’t know where you are today! Rationalisation without this knowledge is fraught with risk of service outages or performance degradation.
You can’t secure, manage, optimise, or validate that you are compliant if you can’t see what is happening across the Enterprise Infrastructure. When it comes to networks, most of us don’t have an idea of which applications are talking to each other, or what circuits and router interfaces they’re using — when servers are communicating, where they are, what their interdependencies are or whether the conversations are even necessary.

Unique insights This network mystique is now being drawn aside by a new service that combines the powerful data collection abilities of HP’s ProCurve Networking switches with the analytical skills of Analytical Applications Limited (AAL) – a leading software and service company that specialises in network business intelligence and chargeback.
Offered by AAL in association with ProCurve Networking, this new service shines powerful lights into the deepest recesses of your network and converts the data it finds into intelligible business terms.
The resulting reports can uncover huge opportunities to save costs and eliminate waste, non-compliance or rogue usage of your infrastructure, while creating a new baseline of how networks, bandwidth and infrastructures are being used by the business. This information can then be used to update the Asset Register and Configuration Management databases that are the cornerstone of IT Infrastructure Library (ITIL) and IT Service Management (ITSM) best practices. It provides the raw material for rich, multi-dimensional enterprise models that provide practical advice on how to run your business more efficiently and it builds into a sound base for management decisions on change.

Simplicity and power While there may be other ways of obtaining this data, using ProCurve switches at strategic points, or throughout the network is less expensive, easier, quicker and more comprehensive than alternatives.
Their suitability for this purpose centres on ProCurve’s use of the sFlow open standard (a method for sampling traffic from across the network). sFlow is an industry standard technology for monitoring high speed switched networks. It gives complete visibility into the use of networks, enabling performance optimisation, accounting/billing for usage, and defence against security threats. As this technology is integrated within the ProCurve ProVision ASIC chips, sampling can be used across the network with no performance degradation.
This produces a comprehensive source of information on the traffic that passes through the switch, tracking conversations from one IP address to another and one server to another, pinpointing their locations and also collecting other data on number of bytes, protocols, time stamps and circuits.
 sFlow Analyser infrastructure
“We have now turned that into a service which can tell you which server talks to which and how the whole thing hangs together,” says Peter Meehan, Director, Business Development of AAL. “If you’re going to move servers, rationalise a datacentre or close it, it can be like a meatball in spaghetti. When you pick it up, and move it, how much spaghetti do you break?
“Now we can tell you which applications talk to which applications, across what circuits and through what router interfaces. We create a multi-dimensional business intelligence model that is of invaluable help to companies wanting to close, move, merge or refresh their platforms.
“At AAL we are agnostic as to where we get our network traffic data from, but ProCurve’s sFlow is ideal for our purposes. We can rapidly and economically put a ProCurve box into a strategic point in the network; or just plug right into an existing ProCurve network. It has inbuilt Internet security and there is no need to impact your circuits by switching on reporting like NetFlow might do.
“We then interpret the data it collects so you can see what is going on and make management decision on what to change or move. It can be surprising what you find – like networks and servers that don’t even belong to you!
“Also, few people know what dependencies have been built into their datacentres around the world. One bank didn’t know its Asia-Pacific datacentre had a dependence on New York until a communication room in New York went down taking Asia-Pacific with it.”
 Understanding the infrastructure complexity
This kind of analysis can also help identify unauthorised or even illegal use by pinpointing unusual address ranges or activities on the network that are worthy of investigation.

Internal proof The AAL/ProCurve service is of particular value to HP departments like Consulting & Integration, providing invaluable baselines for the datacentre rationalisation projects it conducts for customers - one of the most recent being a significant contract to rationalise all 100 of the Times Newspaper’s datacentres across Europe.
HP recently discovered at first hand that the solution really works; on a large-scale Enterprise class network. HP EMEA (Europe, Middle-East and Africa) had embarked on two parallel projects – one an infrastructure rationalisation and the other, a network refresh, led internally by C&I teams. The objective was to not only reduce costs but to maximise the effectiveness of centralised datacentre services over high-speed networks.
AAL had been called in to model the use of the Wide Area Network (WAN) by the HP global business. It was then using Cisco NetFlow traffic flow data and in just seven days, it created a network usage model for the 23,000 servers and applications that support HP’s 142,000 end user community.
However, while it was one thing to model the WAN, the ‘show stopper’was to understand what was happening inside the datacentres, and that is where ProCurve is most effective.
Four key HP EMEA datacentres were involved. The one in Italy had to close and the challenge was to relocate its servers to the most appropriate other centres while maintaining the same performance and service levels.
Simon Newlyn, then head of telecommunications for HP EMEA, saw the opportunity to use smart technology to determine where the servers should go and called in AAL. Accurate traffic flow data was required but Cisco NetFlow was not present inside the datacentre so AAL turned to ProCurve Networking. By drilling into the model, it became clear that five of the top ten Italian servers communicated mainly amongst themselves and predominantly with one of the four new centralised datacentres in the UK. Also, the other servers mainly communicated locally with a minority using international circuits to reach manufacturing and back-office locations.
Prior to AAL being called in, the plan had been to relocate the large servers to one of the datacentres that were geographically close to Italy – either Belgium or Germany. AAL’s analysis of the data sourced from ProCurve made it obvious that the top five servers should be relocated to the UK because conversations from those five servers were mainly generating WAN traffic to the UK. Relocating them close to Italy would have meant that HP would still have had to provision and pay for bandwidth hop from the new location to the UK. Placing them in the UK data centre took virtually all their traffic off the WAN.
AAL’s decision support didn’t stop at eliminating large volumes of traffic. Armed with the knowledge that the top five servers talked to each other, HP consolidated them into one super-computer.
AAL has since worked closely with HP and ProCurve to develop Best Practices and a methodology to create an economic and effective approach to collecting and transforming traffic data into business intelligence. In his new job as manager for solution architecture IT consolidation with HP International Consulting and Integration, Simon Newlyn sees AAL and ProCurve as a vital part of his toolkit. They are used for discoveries, baselines and investigations for IT simplification and datacentre rationalisation projects as well as next generation datacentre virtualisation solutions.

Practical advantages A significant advantage of the new AAL/ProCurve service is that it bridges the gap between business and technicians. The CIO has to talk business language and the way its reports are structured can help him interpret ‘technology speak’ into business language.
It also provides a unique configuration management dashboard that includes both the volumes and costs of the network. There isn’t one database in the marketplace that includes the volumes and costs of networks. The AAL/ProCurve service provides that.
“We interpret bits and bytes to say exactly how much foreign exchanging is going on between divisions in New York and Tokyo for example. We’ve strapped this into real world language and a real world model.”
ProCurve switches are easy to insert into networks under review and because many have sFlow built-in, there is no need to slow your network with CPU intensive protocols like Netflow. Unlike some systems which put in a huge amount of technology to monitor a few applications, they need no additional instrumentation and can model all applications. They are also much more cost-effective than other systems which may be particularly geared for use with WANs.
“Our unique selling point is that we model the clients’ use of their infrastructure, networks and asset more rapidly, economically and easily than any other provider,” concludes Meehan.
|
 |
|