Trading floors are tough and expensive for IT departments to manage. Each trader needs a powerful computer on their desk, connected to a fast network, to drive complex financial applications that require millisecond response times.

When a workstation goes down, the trader is unable to work until IT can get the computer back up and running again. When a technician is dispatched to the desk to fix the problem, it is impossible to tell how long the repair will take. Worst case, a new workstation needs to be configured and installed at the desk which can take hours or days of downtime. The firm loses money every minute the trader is offline. Nobody is happy with this situation.

Today, we can do better. The workstation can run on a hardware cartridge in the datacenter. Not as a virtual desktop, but on dedicated hardware with plenty of performance plugged into a super-fast network. If the workstation goes down, it fails over to another cartridge and the trader is back in business almost immediately.

Beyond reliability, there are major advantages in cost, maintainability, mobility, and disaster recovery achieved by running trader workstations in the datacenter. In this post, we look at how the HPE Moonshot platform can deliver such a solution.

AdobeStock_188823591 (1)

Drawbacks of Desktop Trader Workstations

First, let’s consider some of the drawbacks that come with installing and maintaining powerful computing hardware on the trading desk. The top issues that our clients experience generally fall into these five categories.

  • Downtime
  • Security
  • Maintenance
  • Backup
  • Disaster Recovery

Downtime

Hardware and software crash. When things go wrong, traders need to get back up and running quickly because for every minute they can’t trade, or access analytics, the firm is taking risks, potentially losing money, or missing out on opportunities.

The options for failover on the trader’s desk are limited. That’s the problem. Unless they have a second machine sitting there as a standby, there is no way to quickly get them back up and running without having a technician visit their desk, analyze the problem, and either swap in a new machine or repair the existing device on the spot. Either way, the process of getting the trader back online is too slow.

Security

It is tough to secure hardware that is sitting on a trader’s desk. Data can be copied to a USB flash drive. Hard disks can be removed. Entire machines can be stolen. Sure, there are countermeasures that your firm can take to minimize these risks, but you cannot entirely eliminate the security risk posed by physically storing data on hardware located on the trader’s desk.

Maintenance

Hardware and software need to be maintained. From memory upgrades, to increasing disk storage, to applying operating system patches, to installing new security software, you need to invest resources in keeping trader workstations up-to-date.

But, maintenance is unnecessarily difficult and costly when the hardware is physically located out on the trading floor. Often, a technician needs to physically visit the machine to perform maintenance. This needs to happen off-hours to avoid downtime. Internally developed, proprietary, applications require specialized skills and training to install and upgrade. In a time when IT budgets are under pressure, this maintenance burden consumes resources that could be more productively deployed on other priorities.

Backup

Backup is also more complex, costly, and error-prone when you need to integrate your firm’s backup systems with each individual workstation on the trading floor. In this situation, backup processes compete with business applications for local CPU and bandwidth resources, degrading performance.

As a result, backups may be scheduled for off-hours. So the backed up data may be up to 24 hours old. Can the firm really afford to lose 24 hours worth of local data stored in a trader’s workstation? In many cases, this is a risk that firms take because they don’t have a better solution.

Disaster Recovery

Disaster recovery preparations can be enormously expensive for trading operations. In many cases, a firm needs to lease dedicated floor space, in a remote location, and build out that space with powerful computing workstations on each trader’s backup desk.

The need to support desktop hardware forces companies to lease all this extra space. That’s because there is no way to quickly stand up those local computers in the event of a disaster. Each trader’s machine must be configured, tested, and ready-to-go at a moment’s notice. So, the entire trading floor, complete with the local computers, must be set up and continually maintained.

Given all these drawbacks, it makes sense to consider some alternatives to the traditional deployment of powerful computers to the trading desk. If you could achieve the same performance, wouldn’t it be better to centralize the computing hardware in the datacenter?

New call-to-action

Workspace on Demand with HPE Moonshot

HPE Moonshot technologies provide a viable alternative to desktop computing for trader workstations. The central idea involves relocating the trader’s computing resources to the datacenter, but not as in a traditional virtual desktop infrastructure (VDI) implementation. Instead of sharing hardware with other users, each trader is supported by a dedicated physical machine, running only their applications, but residing in a rack mounted chassis in the datacenter.

Single Tenant Architecture

With HPE Moonshot, each trader runs a dedicated HPE ProLiant m710x cartridge that provides as much, or more, dedicated computing horsepower as a desktop machine. Built on Intel Xeon Server processors with integrated Intel Iris Pro Graphics P580, these cartridges provide 3.7 GHz cpus, up to 64GB dual channel DDR4 memory and dual 10GbE NICs per user.

null

The HPE Moonshot Cartridge provides dedicated HPE ProLiant hardware for each trader.

Each cartridge is housed in an HPE Moonshot 1500 Chassis which can host up to 45 individually serviceable hot-plug cartridges and two network switches in a 4.3U form factor. The chassis is outfitted with a designated switch module that is paired with an uplink module that communicates to your external network on the trading floor.

Although you can have up to 2TB local storage on each cartridge, a better solution is to employ a separate software defined storage (SDS) system. This does not degrade performance, and provides enormous reliability benefits that are discussed below.

On the trading desk, you can either repurpose the existing hardware to provide access to the cartridge and power the monitors. Or, you can replace the existing hardware with a new thin client, like the HP t730 that can drive up to 6 monitors.

 

nullA Trading Desk Powered by HPE Moonshot

Centralizing the trading desk CPUs to the datacenter like this provides a significant return on investment and eliminates many of the drawbacks associated with desktop hardware.

The ROI with HPE Moonshot Trader Workstation

Many are surprised to learn that the cartridges are not terribly expensive when compared with the cost of a high-end desktop workstation. Prices have come down and are competitive, or even lower priced, than their desktop counterparts.

The return on that initial investment comes from the operating expenses saved from not having to maintain desktop computers on the trading floor. Consider how these issues are resolved:

Downtime - Unlike with a desktop workstation, when a cartridge goes down, the datacenter staff can have the trader up and running again on a new cartridge within seconds. The exact configuration of their machine, complete with data that has been backed up nearly in real time, can be restored to another cartridge.

Security - By centralizing the desktop computing and data storage resources in the datacenter, the risk of a security breach is greatly diminished. Employees on the trading floor no longer have access to machines, physical disks, or USB ports that could be used to copy data off of hard drives.

Maintenance - Maintenance is much easier and less time consuming to perform on machines that are physically located in the datacenter. Also, because of centralized data storage, many tasks, like patching, are streamlined. For example, if you have 100 traders running Windows , there is only one “gold copy” of Windows r that all these machines access. So, the patching process only needs to upgrade that one “gold copy”. When the cartridges reboot, they are all automatically running the new version of Windows.

Backup - Backup is simpler and more reliable because you no longer need to backup data from the desktop or compete with trader applications for CPU and bandwidth. Backups can run continuously against the centralized data storage, making a near-zero recovery point objective (RPO) possible.

Disaster Recovery - Dedicated floor space leases may no longer be necessary once you have moved to an HPE Moonshot solution. A failover workstation environment can be set up on a rack at a remote datacenter. Should a disaster strike, traders can set up operations in a hotel with a laptop or thin client and monitor.

For all these reasons, a significant return on investment is achievable by moving to trader workstations with HPE Moonshot. But, there are other opportunities as well.

Who Can Benefit from HPE Moonshot?

Financial services firms can benefit from HPE Moonshot if they support operations requiring reliable, high-performance workstations with real time data access. This includes:

  • Large Financial Institutions
  • Hedge Funds
  • Quantitative Analytics Firms

But, not only employees who need powerful financial analytics can benefit. Any employee that needs dedicated hardware to run resource intensive applications in an individual environment can benefit.

Software Developers Also Benefit

Software developers are a good example of another category of user that can benefit from HPE Moonshot. Developers run complex, memory and cpu-intensive applications on personal machines to that they can write and test code in an individually tailored environment. Software running on the personal machine of a developer may include database engines, web servers, IDEs, device simulators, etc. When such an environment goes down, it can take days for the developer to rebuild. So, the type of failover provided by an HPE Moonshot system is a huge benefit.

Offshore and temporary employees are good candidates for this type of solution. You can provide these developers with secure, high performance, resources and a seamless user experience.

There are many cases where HPE Moonshot can improve performance while simultaneously reducing operating expenses associated with high performance workstations. And, with support from HPE, you can even finance the initial acquisition of HPE Moonshot technology out of operating expenses.

IIS - Your Partner for Trader Workstation Solutions

International Integrated Solutions (IIS) is a managed service provider and system integrator with deep expertise in financial services and HPE Moonshot. IIS is a distinguished HPE partner, winning HPE Global Partner of the Year in 2016 and Arrow’s North American Reseller Partner of the Year in 2017.

As your systems integrator, IIS brings decades of experience supporting high performance financial applications. This experience makes them well suited to helping you realize the benefits HPE Moonshot. In particular, IIS can help with:

  • Sizing - understanding your workloads and specifying hardware that can deliver the performance you need at a reasonable cost.
  • Migration - helping you move your applications onto the Moonshot platform.
  • Backup Solutions - integrating trader workstations into a comprehensive backup solution to minimize recovery time.
  • Disaster Recovery - designing a cost-effective solution to meet you business continuity requirements.

Move Your Trader Workstations to the Datacenter Today

Particularly if you face the need to refresh aging desktop hardware, you should consider migrating to HPE Moonshot. Before you rollout new desktops, look at the reliability, maintenance, and overall cost advantages of centralizing your desktop computing resources in the datacenter.

 

Jose Restrepo

Written by Jose Restrepo