As packets of information travel from, say, New York to Tokyo and London, a company can lose money because the data reaches its offices at different times. That, say GigaSpaces Technologies and Solace Systems, is the problem they are seeking to address with jointly developed messaging technology.
The product, which is designed to speed up communications and synchronize data from applications running in separate locations, is particularly helpful, according to the vendors, for financial firms that need to coordinate interactions between systems running in multiple clustered computing environments, or in cloud computing.
Ottawa-based Solace makes hardware that accelerates the flow of messages by eliminating the system interruptions and other traffic obstructions that come with software-based messaging products, said Chris Pittaway, sales director for Solace’s Tokyo office. “And the other benefit,” he added, “is that the latency that you do get is very consistent–you don’t have the outliers that you would get with a traditional software-based solution.”
By integrating Solace’s hardware with the GigaSpaces eXtreme Application Platform, or XAP, firms can move data across wide area networks more quickly and efficiently, say the companies, allowing for distributed data mirroring across distant geographies. Neither company has customers using the combined technology, which is now available, but there are pilot projects under way, according to Pittaway.
The question may be, is this a real problem, or technology looking for a market? “It could be a bit of both,” said Neil Katkov, a Tokyo-based analyst with research firm Celent. “But it’s a trend that we’ve seen, a response to the growing complexity of the brokerage technology space.”
Part of the problem, noted Katkov, is the widespread use of algorithmic trading strategies and automated trading systems. “This is producing a lot of data issues for companies,” he said. However, he added, “it’s hard to say how applicable or useful a financial services firm is going to find a particular vendor’s products.”
A problem with some applications is that they are designed to wait for confirmation that a message was received before sending a second, said Shawn McAllister, VP of architecture at Solace. On a local network, where computers are physically close together, that does not add up to significant delays. “But when you add in the distance between New York and London, that could generate 50 milliseconds of latency per message per round trip,” McAllister said.
Solace sends messages one right after another–resulting in hundreds of times more messages if long distances are involved. But the product also checks that messages are delivered and includes management controls to help diagnose problems. “This is a much more efficient use of existing bandwidth,” asserted McAllister.
A typical asynchronous communication approach–where one side sends out messages without pausing for a response–can result in lost data, said Shay Hassidim, CTO of GigaSpaces, which has offices in New York, San Francisco, London and Herzliya, Israel. The GigaSpaces-Solace approach breaks up the journey into several segments, said Hassidim, with a checkpoint at the end of each to check the integrity of data, and buffers to keep the information moving as quickly as possible. This, he said, creates a “totally reliable communication chain between source and target.”
A typical application of the product, say the companies, would be in backup data warehouses for disaster recovery purposes, or in data repositories in multiple locations for quicker access by far-flung branch offices.