Legal Technology Journal

How your firm can avoid a nasty shock when the next big business application is selected.

The scenario is a familiar one. Key people from the business, analysts and a large group of IT professionals spend weeks comparing two or  three candidates for a new CRM, DMS or finance system. It was a close-run thing, but Mivvi 2010 wins the day: ‘The users love the interface and it’s £150,000 cheaper.’ 

Roll-out to head office goes smoothly, everyone’s happy and the project is hailed as a great success. Then comes deployment in Paris and the problems begin. From day one, the French users complain – not just about the new application, but also about the sluggish performance of all systems. After some detailed analysis two things are discovered: first, Mivvi 2010 is heavily impacted by any delay on the wide-area network; and secondly, the application is very bandwidth-hungry. The bottom line is that not only will it be a struggle to deliver satisfactory response times, but an £800,000 upgrade to the global WAN is needed to accommodate the extra load. What’s more, the lead time on some circuit upgrades is three months and that means the project will slip, incurring more cost. Suddenly Mivvi 2010 is not looking so good.

In the world of applications, it can seem like you’re spoiled for choice. The number one factor in making that choice will always be functionality. It doesn’t matter how fast an application responds or how little load it places on the WAN, if it doesn’t meet the needs of the business, it’s useless. Senior IT consultant at Advance7 Addy Kedia has helped a number of organisations run product selection tests. ‘It’s important to make sure the application does what the business wants, and not the other way round,’ Addy explains. ‘That’s why proper functional testing is crucial. Check that the product answers your requirements.’ Given the broad choice of applications, it’s rare that only one will meet the business needs, and this is the time to consider some of the other issues.

Some things to consider

There are few things that will upset busy business people more than giving them a new application that delivers slow or erratic performance. It’s also unpopular to spring the need for a multi thousand pound upgrade onto the project manager during roll-out. Ideally, you need to know two things before you even choose your new CRM, ERP or DMS system:

• How will response times for each product measure up when run out of your far-flung offices?
• How much bandwidth will each product need, and hence what are the upgrade costs likely to be?

For an informed decision, you need this information right up-front. Many organisations will address the performance and upgrade issues prior to deployment, but few will tackle it during product selection, and there are two good reasons: cost and time.

Trying before you buy

The standard approach to gaining performance and load information is to build a test environment with load generation capabilities, determine the current transaction rate and mix, write complex load scripts and run a raft of tests. This is great for pre-deployment testing but not ideal at the product selection stage as the whole process has to be repeated for each product being considered.

Paul Offord, MD of Advance7, talks about the slimmer process they have developed to make pre-selection testing a viable proposition. ‘The process can be used in a proof of- concept environment without costly load generators and uses WAN simulation and mathematical modelling techniques
to deliver:

• the response times to expect at each remote office; and
• the comparative load generated by each product.’

Offord adds: ‘Heavyweight pre-deployment testing will take between 15 and 20 days per application. The slimline approach we’ve defined gets the comparative figures needed in three or four days.’

Whilst application functionality remains the number one aspect to consider when selecting a new business application, performance at remote sites and upgrade costs run a close second and third. It’s quite feasible to use a lightweight process to run comparative tests, and with a slight tweak even include test for the effectiveness of those wizzy new WAN acceleration products. At the end of the day, demanding side-by-side comparative testing means a firm can be confident that its investment is a wise one. So much for the theory but how does this work out in practice?

Testing in the real world

Simmons & Simmons enlisted Advance7 to provide impartial evaluation of certain software products as part of the firm’s broader IT upgrade. This upgrade entails a total rebuild of its IT infrastructure and includes a new global document management system, a new version of Microsoft Office, and Outlook and Exchange for e-mail.

One driver to this IT upgrade, codenamed Programme Vanilla, is to meet specific needs created by Simmons & Simmons’ international span and the globalisation of businesses more generally. This has caused a rapid rise in cross-border work where information and case documentation has to be shared in multiple offices around the globe. Simmons & Simmons’ legacy DMS did not allow simple sharing of documents around the world. It was also limited in that e-mails and other data types could not be stored in the DMS. Following a formal tender process, Simmons & Simmons enlisted the services of Advance7 to assist in the evaluation of two leading products; one from Hummingbird (now Open
Text) and another from Interwoven.

The evaluation was based on three broad criteria:

• Functionality – would the DMS provide the facilities that Simmons & Simmons needs on a global basis?
• Performance – would the performance be acceptable at all offices across the globe?
• Load – how much load would each DMS network infrastructure put on the IT network and bandwidth capacity?

Simmons & Simmons set up two test environments – one for each DMS solution. In each, three virtual offices were constructed: London, Hong Kong and Shanghai. London and Hong Kong were data entre locations, and Shanghai was the farthest-flung office on a low-speed link. This combination enabled measurement of client-to-server data flows and server-to-server replication traffic. Advance7 supplied and configured WAN simulators to accurately recreate network delays and  bandwidth constraints.

Simmons & Simmons test users then assessed the functionality and performance of each system while Advance7 specialists controlled the test environment and made diagnostic measurements.

After one month, the test results distilled to provide three important pieces of information:

• Functionality – each system had its strengths and weaknesses but one more closely matched the needs of Simmons & Simmons and therefore would deliver the greatest benefit to the firm
• Performance – given the proposed configuration, one of the DMS solutions would not meet the response-time needs in Shanghai and other remote offices
• Load – the additional network load and maximum tolerable latency was determined directly for the three virtual offices and extrapolated for the remaining 19 offices.

Ultimately, the information from the project enabled Simmons & Simmons to back up its own internal findings and select the correct DMS to meet its requirements based on solid evidence. According to Abby Ewen, IT director at Simmons & Simmons: ‘Provision of a robust, global DMS system is of critical importance to our business and enables the firm to provide its clients with cross-border advice on complex legal situations. However, the DMS does not stand alone, it impacts our network, bandwidth usage and a multiple of other IT technologies employed at the firm. With independent and robust test analysis of the available technologies in our specific test environment, Advance7 provided results and 100% impartial advice that gave us the indisputable proof we needed to make the right investment decision.’

Simmons & Simmons selected the Interwoven Worksite DMS as its global repository.

Charles Christian, Editor of Legal Technology Journal