The Case for Wide Area Network Optimisation Solutions in Zimbabwe: Part 1

Prince Dube Avatar
Server Room

Server RoomThis is the first article in a two part series that discusses the challenges faced by Zimbabwean organisations that have branch networks and are constantly seeking redress to their problems. The first part will discuss the overview on challenges faced by most network managers and how WAN optimization solutions can address these issues and the second part will drill down into WAN optimization options available to the Zimbabwean market.

Background of challenges faced by Zimbabwean corporates

We would have loved to discuss in depth the WAN connectivity options available to Zimbabwean based corporates but merely mentioning them throughout this discussion should suffice as most IT people in these organizations have already conducted extensive research on the technologies and probably have already implemented solutions that are still giving them sleepless nights.

IT administrators in organizations with branch networks and a mobile workforce are faced with the perennial challenge of managing a variety of traffic types that traverse their networks. Regardless of the technology chosen (be it VSAT, Frame Relay, WiMax, 3G, leased line or indeed optic fibre), response times of critical applications are a pain somewhere for most managers.

How many of us have tried to change from leased lines to VSAT or vice-versa, and even upgraded from a 512kbps link to a 1Mbps link (and started paying double for that matter) but still cannot get to the bottom of the issue? Bandwidth costs represent a significant proportion of operating expenditure for wide-area data networks but the cost is not the only consideration. Matching the allocation of WAN resources to business needs is also very important.

And as resources are increasingly centralized, minimizing the effect of latency on application response times is becoming a critical requirement. In addition, virtualization and new application environments, like cloud computing and web services, can put unexpected strain on the network.

Different types of traffic and IT infrastructure present both difficulties and opportunities for improving the response times of essential applications. For example:

  • Traffic that isn’t time-sensitive, like e-mail, backups and personal web access, can swamp WAN links, leading to slow response times from business-critical applications.
  • Applications, such as Microsoft SharePoint, that make extensive use of dynamic content can swamp WAN links while delivering poor end-user response times.
  • Global centralization of branch office servers and data centers can expose latency-sensitive protocols, again leading to slow response times.
  • File transfers, operating system and patch distribution and similar applications, such as antivirus updates can quickly saturate WANs.
  • Repeated transmission of the same, or similar, files, objects or data patterns can create opportunities for data compression.

Since optimizing overall application response times is a requirement for many organizations, this discussion summarises solutions that address the common need to make more efficient and effective use of wide-area connections, regardless of the type of traffic or application.

The predominant need is still to optimize the connection between users (both in remote branch locations and single remote users) and centralized ICT resources. However, we are beginning to see the emergence of the need to optimize connections between data centers. We are also seeing early signs of the need to optimize traffic to mobile devices due to the emergence of technologies such as 3G and 4G by Econet and Ecoweb.

The development of the application acceleration market has been driven by customer demand for highly integrated solutions that employ a wide range of techniques to optimize network traffic, and that offer scalability and fault tolerance. Vendors in the WAN optimization space such as Cisco, Juniper, Packeteer and Riverbed initially addressed either the traffic shaping/quality of service (QoS) market, or the compression/caching market. These two segments have now largely merged, with most products supporting both sets of capabilities.

While these capabilities address the problem of inadequate bandwidth which is slowly becoming a thing of the past with the emergence of optic fibre networks in Zimbabwe, network latency (especially on VSAT links) has become a limiting factor on remote application performance. We therefore see an increasing need for both generic and application-specific optimizations to mitigate the impact of network latency on remote application performance. While high speed networks are slowly becoming available to Zimbabwean organizations for connectivity in the main centres such as Harare and Bulawayo, we may still be a long way in realizing an optic fibre connection in areas such as Nyanga, Chimanimani, Chiredzi or indeed along the Great Dyke-one of the most economically viable sections of this country.

In addition, the following WAN Optimisation Controller (WOC) product trends are emerging:

  • In branch offices, the capabilities of WOCs will evolve to the point where they can support server-less branch operations, also described as Branch Office Boxes (BOBs). This will require (as a minimum) the addition of supporting features including Dynamic Host Configuration Protocol (DHCP), Domain Name System (DNS), active directory caching and print serving.
  • Customers often need to maintain one or two key applications in the branch. BOBs are now leveraging operating system capabilities, including virtualization, to host one or more applications on the BOB hardware.
  • There is an increasing focus on security especially with the increased opening up of the Internet space to Zimbabwe (what with Powertel announcing their connection through Botswana and other providers making conceited efforts to connect through Moazmbique and South Africa) — including the acceleration of encrypted protocols such as Hypertext Transfer Protocol Secure (HTTPS) and the security of data stored on WOC systems.
  • As basic acceleration capabilities mature, we expect a resurgence of interest in visibility and control, both as a means to demonstrate WOC effectiveness, and as a bandwidth/response-time planning tool. Providing application and user performance measurement and SLA reporting are key emerging requirements for WOC equipment. As Zimbabweans we need to slowly move away from focusing on products and more towards service management.

Market Definition/Description

Having mentioned WOCs in the paragraph above, it is imperative that we revisit this and provide a detailed description of this animal for the benefit of the reader.

A WOC (WAN Optimisation Controller) is customer premises equipment that is typically connected to the LAN side of WAN routers, or is software integrated with client devices. WOCs are deployed symmetrically — in data centers and remote locations — and improve the performance of applications that are accessed across a WAN. They address application performance problems caused by bandwidth constraints and latency or protocol limitations. The primary function of WOCs is to improve the response times of business-critical applications over WAN links, but they can also help to maximize return on investment in WAN bandwidth, and sometimes avoid the need for costly bandwidth upgrades. To achieve these objectives, WOCs use a combination of techniques, including:

  • Ensuring fair access for mission-critical applications during periods of congestion by prioritizing business-critical traffic, through QoS policing and traffic shaping, for example.
  • Minimizing the effects of network latency using methods like protocol- and application-specific optimization.
  • Reducing the bandwidth required to transfer WAN traffic; by compressing it, for example.

Prince Dube is an ICT practitioner based in Harare, Zimbabwe. Prince has been involved in the goings on of the local ICT industry for over 10 years and currently works for Axis Solutions. He can be contacted directly on pdube@axissol.com.

2 comments

  1. Christopher Muyaruka

    I used to work for a large NGO with 6 remote sites and I was at constant logger heads with Finance in justifying the cost of our point to point links, I think service providers are also to blame for not clearly spelling out parameters like contention ratios and latency on links they provide – at one time I logged a call with one of th providers for a slow link, the response of the technician was mind boggling – he just told me his job is to check if there is a green light on the customer equipment…..

  2. Prince Dube

    Well said Christopher,the area of customer service is one that seriously requires a paradigm shift and we hope with the coming in of international investors competition will force service providers to treat their customers with respect. While that may be there is still the issue of inherent latencies in these links that may only be solved by managing the applications and services so that they run better on the same links.

2023 © Techzim All rights reserved. Hosted By Cloud Unboxed