ParaScale: Building parallel and scalable storage technology in the cloud
Date: Saturday , May 02, 2009
Wouldn’t it be nice to host our own Web-scale file system to handle all the data that it processes and stores? As Web-scale computing and the needs of enterprise storage and security grow, many enterprises are thinking of this possibility. Though large corporations like Amazon and Google have already done it, commoditised product version of this was perhaps unheard of. This became an opportunity for California based Parascale, which provides storage in a networking "cloud" for digital content providers. ParaScale provides a software application than runs in user space on multiple standard Linux servers. The result is a highly scalable self-managing storage cloud, with massive capacity and parallel throughput.
Cloud computing is grabbing headlines as more and more enterprises are turning to this emerging technology as a way to enable flexible computing power over the Internet. In 2008, many companies were wary of the risks and vulnerabilities of participating in the cloud computing model, and although the buzz was high, adoption was feathered. The economic downturn and the addition of private cloud solutions to complement public offerings are creating an environment that enables incremental adoption of cloud storage on a very broad scale. The company CEO Sajai Krishnan eyes a fair share in a roughly $15 billion estimated market.
ParaScale was built to address the needs of emerging markets while being able to scale to hundreds of nodes in a loosely coupled architecture similar to the Google file system. Krishnan points out that many organizations are considering public and private cloud based storage. Building storage clouds is becoming as simple as installing a new application on your laptop. This is enabling service providers and the enterprise to embrace this technology with minimal effort. Cloud storage can start small and scale-up as needed. Organizations are no longer over-building to address the potential for rapid growth. Instead the drive is to put in place an architecture that is extremely flexible and that can scale on demand using commodity hardware and standard client access. Moreover clouds are designed to be self-managing and don't require heavy IT manpower. Storage tiering, provisioning, and data movement are time-consuming tasks that are automated in cloud storage. Storage clouds can be tuned for specific uses or applications. For example, clouds can be tuned for archival very cost-effectively, or for streaming media performance. http://www.parascale.com"
The ParaScale Cloud Storage (PCS) software runs on a Linux server functioning as a metadata controller, and on standard Linux servers connected by gigabit Ethernet. The file storage on these servers is turned into a single global name space with access via NFS, WebDav, HTTP or FTP. ParaScale says PCS is an object-based cluster file system. To add more capacity you add another Linux server which also adds more I/O bandwidth to the overall cluster as the nodes operate in parallel.
PCS is a loosely-coupled cluster, in the sense that there is no global cache across the cluster and no-high-speed interconnects such as InfiniBand or 10GbitE. ParaScale is pitching it for tier 2 (non-transactional) storage, providing good access to many, many files stored on hundreds of nodes in parallel. The software can migrate files across cluster nodes for load-balancing purposes, which can solve the problem of limited bandwidth to individual nodes throttling performance.
Companies can keep adding as many servers as they need, with each one acting as a redundant node. The software runs on the cluster as whole, treating it as one giant file system. This creates private cloud storage that companies can host themselves inside their own firewalls. ParaScale CEO Sajai Krishnan says “ParaScale's approach to data storage is transformational and will help IT managers address increasing file storage costs as well as capacity growth. According to him, Cloud computing is driving the next generational shift in IT and as we move forward cloud storage will be a major way that storage is consumed. customers can expect to pay about $1 per gigabyte, depending on their server costs. Several customers have been beta-testing the software, for both private and public cloud applications, such as Blue Coat Systems, the Stanford Genome Technology Centre and Sony Pictures ImageWorks.
The kinds of applications that make sense for ParaScale include video hosting, applications that crawl the Web and create huge log files, or corporate databases that are simply getting out of hand. Steve Kowalski, senior systems architect at Sony Pictures Imageworks said, “We have been able to leverage commodity economics in all parts of our infrastructure, but storage has been the holdout. After realizing benefits across workstations, render farms, and operating systems, we have the opportunity to see the benefits of commodity hardware in storage.”
ParaScale Cloud Storage software Version 1.2 is available for immediate purchase. Pricing is based on the physical capacity within the cloud, with deployments starting at a list price of $1.05 per gigabyte.
Currently cloud computing represents a sea change in the way IT is architected. With enterprises under severe economic pressure to innovate and cut costs, the cloud storage alternative is one that seems to makes sense. It is pegged to be a $40 billion by 2012. Krishnan finds that the immediate market in the streaming media space is about a billion dollars worth and it will ramp up by about 100 percent every year, and a much larger archival market which though not growing in dollar terms, is going through a radical refresh. With less than 50 employees, the company has raised $11.37 million from venture capital firms Charles River Ventures and Menlo Ventures till now.
ParaScale may just be at the tip of the iceberg as startups and established storage vendors see the profit possibilities in clearly conceived, secure cloud storage technologies. ParaScale, the name being a combination of “parallel” and “scalable”, is on a path to secure, convenient, scalable and available storage, that has never been clearer. ParaScale plans to sell software, to service providers who can deploy public clouds to deliver public storage services over the internet,or to enterprises who can buy it to create their own private clouds behind their firewall. No doubt, the storage marketplace has gone through several major shifts over the past twenty years. So, with cloud storage, it looks like we may be seeing another shift – and Parascale will now have the resources to become a leader in the space.