At first glance, private cloud computing and hedge funds would appear to be a marriage made in heaven, especially for small-to-medium sized funds. The financial, operational and time-to-market advantages alone would seem to carry the day. But, before they tie the knot, hedge funds want providers that specialize in their business needs.Such is the case for a start-up hedge fund in San Francisco, Resultant Capital Partners, which is using the AbacusFlex private cloud services of the Abacus Group, also based in the City by the Bay.“First and foremost, we selected Abacus because of their knowledge of the hedge fund industry,” says Michele McGovern, a partner and chief operating officer (COO) with Resultant. During her first months with the fund, McGovern says she focused on finding the best-in-class IT providers such as research systems and order management systems (OMSes). She also honed operational business models, and established lines of communication with prime brokers and the fund administrator, Citco Fund Services. She also spent time considering whether or not Resultant should set up its own infrastructure.
Bob Guilbert,
managing director,
Eze Castle Integration
“When we were looking at the possibilities for putting an infrastructure in place, one model is to buy the servers, run your own network, implement your own applications, and basically be responsible for your own data center. We have a lot of experience doing that and, at the previous hedge fund, that’s what this team did,” says McGovern about Resultant’s founders Jeffrey Skelton, managing partner and CEO, and Michael Henman, partner and head of business development; Skelton and Henman are the former founders of Symphony Asset Management. (See the Q&A about the founding of Resultant)But, over the past decade, the elements of cloud computing—the virtualization technologies, security, reliability and performance have evolved to the point where funds like Resultant are compelled to consider the private cloud option. McGovern was looking for a provider with the extra hedge fund infrastructure expertise.
“We also selected them because of the quality of the infrastructure that they have in place—the disaster recovery back-up in Philadelphia and the production data center they have here in Oakland,” she says. In addition, Abacus staff members dedicated to the West Coast are based in an office a half-block away from Resultants’ offices.
“When we need them to work with us on special projects or implementations, they’re here in our office,” McGovern says. “The client service model from our perspective is very important. When you need some help or you want to talk about a new project coming online, building out more of our research environment, they come here and collaborate with us. They’re very hands-on. … That was one of our key decision-making criteria.”
New Attitude
Abacus manages security issues, connectivity to prime brokers and fund admins, hosting for OMSes and applications, and scalability.
“Over the past nine months, we’ve gone back to Abacus three times and increased our needs and requirements in terms of data storage capacity and asked them to host additional applications for us,” McGovern says. “The benefit is that there isn’t any lead time. We don’t have to order servers and wait three months for them to come on line. Abacus already has this infrastructure in place. We can just work seamlessly with them and get this done in a matter of weeks as opposed to months.”
Abacus hosts Eze Castle’s middle office software system for Resultant. The fund oversees the trade allocations and the sending and receiving of trade files to its prime brokers and to Citco. Resultant also coordinates with its prime brokers and Citco for any trade breaks, settlements and allocations. “Eze Castle has its team monitoring the send/receive files,” McGovern says. “The back office is done by Citco with us doing the shadow work.”
Most of Abacus’ customers retain networking for local area connections as well as routers and switches and firewalls, says Chris Grandi, founder and principal of the Abacus Group. “We actually put hardware on site and it’s really a caching server.” The caching is another layer of protection against multiple points of failure.
“We don’t have to order servers and wait three months for them to come on line. Abacus already has this infrastructure in place. We can just work seamlessly with them and get this done in a matter of weeks as opposed to months.”
— Michele McGovern, COO with Resultant Capital Partners
Like other cloud providers, Abacus also handles data aggregation, which is a major challenge for firms that use a lot of data—the quant and black box funds, which need lots of market data for their pricing models.
“The challenge is that when you internalize your IT infrastructures—there were always limitations to how much data you could bring into your environment,” Grandi says. “If you’ve reached that capacity, then you have to buy additional hardware and additional networking equipment. … We just throw more processors, memory and storage space at the problem.”
But not all start-up hedge funds will automatically jump onto private clouds despite the likely financial and operational advantages. A great deal rests with the attitude of the portfolio manager and the principals at the hedge fund, says Bob Guilbert, managing director of Eze Castle Integration, a provider of cloud and other hosting services for hedge funds and investment management firms. (Eze Castle Integration is separate from the company that provides the Eze Castle transaction software.)
“We see still see new clients—start-ups—that want to own everything,” Guilbert says. “We also meet people at other funds—whether it’s the CEO, COO or CFO—saying that they do not want to be on the treadmill.” They want to sidestep the ongoing expenses of data center support coupled with hardware and software maintenance and upgrading.
Yet, if a fund turns to a private cloud, how does it use IT to its advantage in a highly competitive market?
Chris Grandi,
founder & principal,
Abacus Group
One answer lies with the openness of the cloud provider platforms, says Alexandra Lin, a London-based analyst with market research firm Celent’s security and investment practice. Clients are pushing private cloud providers to support other vendors’ applications and systems, which will help hedge funds maintain their IT advantage.
Specialization will underscore the IT advantages that cloud providers can offer, Lin says.
“The specialists will be better equipped to ensure that hosted management of hedge funds, proprietary research, pricing, performance systems, order management systems and FIX connections will be proactively monitored and managed,” Lin says. The cloud provider also needs expertise in hedge fund applications, database offerings, pricing fees and integration points to prime brokers and fund administrators.
Early on, Resultant ruled out the general, “Amazon-like” cloud providers that allow firms to rent servers inexpensively via the cloud. “The model is not oriented toward the specific hedge fund requirements,” McGovern says.
Those specific requirements are illustrated by Resultant’s need for email archiving. “There are two service providers that do that and we wanted to know that Abacus could connect with them and that they understand what the requirements are from an SEC perspective,” she says. “The benefit of working with someone in our space … means they’re keeping in tune with the same rules and regulations that we have to.”
Keeping the Faith
However, specializing in private cloud services took a lot of courage three years ago when the hedge fund industry was in dire straits Grandi can attest. The former president of Eze Castle Integration launched the company amid a great financial crisis.
“It was probably the best time and the worst time to start a company,” says Grandi, who adds that he sidestepped renting office space and other costs to put a limit on Abacus’ outstanding liabilities. Despite the external signs, Grandi and his fellow pioneers held to their belief that they would succeed. “We’re all hedge fund guys so if we doubted doing this, we’d have to doubt our entire careers because we’re tied to hedge funds.” Keeping the faith helped.
“Through 2008 and 2009, most hedge funds did not spend money internally on their infrastructures because they didn’t know how long they were going to be around,” Grandi says. But, by the end of 2009, firms got more comfortable and many realized that they were going to weather the storm.
Business has picked up for Abacus over the past 18 months and it has more than 80 hedge—with 20% of them new funds that launched over the past two years months, Grandi says. The rest have assets under management ranging from $1 billion to $5 billion.
“Through 2008 and 2009, most hedge funds did not spend money internally on their infrastructures because they didn’t know how long they were going to be around.”
—Chris Grandi, founder and principal of the Abacus Group
Eze Castle Integration’s Guilbert reports that he has seen a surge in the use of private clouds among hedge funds. “We’re seeing at least 30 percent of our business tracking toward the cloud,” he says. He expects the pendulum to keep swinging in favor of the cloud—three months from now demand could be 50 to 60% trending toward the cloud.
The trend may be related more to what a cloud provider can do rather than a knee-jerk reaction among funds after the financial crisis to take the most cost-effective route to an infrastructure, Guilbert says.
“I think with the financial crisis there was a higher demand on how capital was being used and how investors in hedge funds are being protected,” Guilbert says. “More questions are being asked about how a fund operates, how the firm protects the data, the applications, and how they will insure that trading continues in the event of a disaster.” On that score, cloud providers can offer disaster recovery fairly seamlessly, he says.
What the Cloud Can’t Cover
Despite the fact that private clouds are becoming building blocks of corporate infrastructures, hedge funds should think twice about putting all of their application into a cloud environment, says John-Peter Lee, managing member of MetaFrame Technology Solutions, a consultancy that helps small-to-medium fund managers build their infrastructures and find outsourcers.
“Applications that need to take advantage of 100% of any computing resource—CPU, disk or network throughput—are best left on a dedicated, unvirtualized server,” says Lee, who founded MetaFrame after serving as the chief technology officer (CTO) at hedge fund AM Investment Partners. Trading algorithms, real-time risk analytics and high frequency trading might be best left out of the cloud. “A cloud provider familiar with the types of applications a hedge might employ can better guide them when formulating a private cloud strategy.”
For many firms, the advantages of a private cloud infrastructure do not outweigh their fears about client data residing in a remote location. It’s an issue that can divide hedge fund managers.
Once firms seriously consider and then understand the concept of cloud computing, they get comfortable quickly, McGovern says. Another school of thought argues that they are uncomfortable unless they own and manage the servers from their offices. “I’ve had a number of those discussions with people at other hedge funds … If you sit down and think about what you’re trying to do with your business, and what your scalability requirements are, who in your firm is an expert in managing all of that? Because it isn’t just buying the servers, it’s the connectivity to multiple prime brokers and fund admins.”
On the issue of data security, McGovern points out that hedge funds work with prime brokers and fund admins. “Your information about your clients and your portfolio is already residing on servers that are not in your offices,” she says. “My advice would be to look at it from that perspective.”
McGovern’s own perspective is cloud providers are on the same level as fund administrators, prime brokers and broker/dealers that have to conform to rules and regulations governing trading interactions.
“Those same rules and regulations apply to Abacus,” McGovern says. “We look at Abacus as experts in infrastructure technology and networking—that’s all they do. They’re dedicated to it and they apply it to the hedge fund industry.”
It’s not a given that all hedge funds will put their cloud providers on a par with prime brokers and fund admins, Guilbert says. It’s also true that private cloud arrangements may not last forever. “If you start to look at larger organizations that have staffs greater than 50 people, the trade-off between going to a cloud on a per user, per month basis versus buying the infrastructure is something the firm needs to assess,” he says.
At some point, all firms need to decide if it’s more cost effective to be in the clouds or to build their own infrastructure. A great deal appears to depend on the AUM level that a fund achieves.
“Once you get up to $5 billion, and we do have clients with AUMs in excess of $5 billion, there’s a certain element of control and the expense to internalize it and they might actually save money and have greater control over their environment,” Grandi says.
Of course, the opposite is also true—a fund could run into trouble and leave the cloud for different reasons. “In our short time, we’ve already had two clients at the $1 billion level close their doors,” Grandi says.
A move off the private cloud for Resultant would depend upon the growth plans of the firm and how satisfied it is with its service, McGovern says. “If our needs were to grow such that the cost-effectiveness of it—meaning data storage costs, service and maintenance costs, and so forth—was at a point where we couldn’t mathematically justify it, then maybe we’d consider bringing it inhouse,” she says. “But I will tell you that our first choice is going to try and work with Abacus to not have to get to that point.”
For the moment, Resultant has no plans to leave its private cloud.
“Our open discussions with Abacus have been, ‘We’d like to be able to scale our business with your business.’ And as long as we have an open dialogue to do that, then I’m quite confident we can figure out a method that works for everybody.”
Need a Reprint?
Leave a Reply