Edu-leaders

Newer courses, an expanding base of demanding students and digital learning—today’s institutes face multiple challenges. To cope, these hubs have to invest in providing provide an international experience to students; 24/7 Internet connectivity, accessible learning material (in form of e-books), e-space within the campus IT network, access to performance and attendance-related data, e-trained teachers and a culture of research through repositories…the list is long. A substancial IT infrastructure is needed to enable this. Well-equipped computers, high-speed networks, Internet and the latest software, are all must-haves.

With all this informative data, storage becomes a key issue. Institutes are still not paying enough attention to this factor—as long as they have big, fat hard drives built into computers and servers. Imagine being able to pool all data generated by staff, teachers and students into an intelligent repository. A praiseworthy lesson plan by a teacher or a project report by a student is often lost in a sea of hard drives when a new computer or server replaces the old one. Most institutes are not cognisant of the volume and quality of information that drains out from the “institutional ecosystem”, as each student, or teacher, moves out. Is there a way to preserve the work they might have done at the institute?

What is the usage of storage solutions in the education segment? We asked Anand Karapurkar, the director and founder of Infobahn Technologies, a company delivering IT-solutions to institutions. He says, “The storage market for this sector has not evolved. Reasons are the cost, which has come down drastically at an enterprise level, but is still high for the education segment. Also, the data generated by institutes are mainly student-related data with a short lifecycle. Though institutes are turning into smart campuses (with high-end IT infrastructure), few have created repositories that allow live access to students.”

However, the statement is not true for all. According to Springboard Research’s August-2009 report, “Inside the Campus: IT in India’s Education Sector”, Storage Area Network (SAN), Wireless LAN (local area network), and ERP (enterprise resource planning) are the three most popular IT-solutions adopted by Indian institutes. But, no two institutes have the same storage requirements. And there is no one-size-fits-all solution. While there are institutes working with enterprise-level storage technologies, there are others happy with their PC-level storage.

How do you then choose the right storage setup? And what factors should you consider while making a choice? How do you arrive at the right storage capacity that is capable of supporting new applications needed by a university? Let’s explore.

Different Needs
All institutes are not alike as far as vision, number of students and courses are concerned. In spite of being part of the same industry, requirements are different.

IIM Shillong runs enterprise resource planning (ERP) software, so there is a lot of online data generated, says professor Basav Roychoudury, the hub’s IT head. The institute’s servers have a dedicated 6 terabyte (TB) space for library purposes as well. According to Roychoudhury, “Our storage backbone is driven by NAS (Network Attached Storage) servers. Each server comprises a group of 20 disks with 140GB and are equipped with RAID level-5 technology to enable data redundancy. To regulate admissions, maintain student records, examinations and running of classrooms we have allocated certain disks on the servers. Backups are kept on tapes that are stored in a fire and electromagnetic resistant environment. The institute has been using this configuration for the past one-and-a-half years.”

He adds, “To select the right kind of setup, the first criterion is to arrive at a capacity (for your institution). This capacity should be such that it is able to support added applications that may be required in the next five years. While choosing a storage setup, look for cost benefits, disk speed and backup server. Conduct a survey for the best deal. Bargain with vendors. Keep space for the operating system that will be running, and get an idea of the amount of data will be generated according to the number of students and multiply that with the number of years that the setup will be used, with enough redundancy built in. For technical reasons the entire disk space cannot be used and for commonly efficient running a maximum of 70 percent to 75 percent can be.”

“While choosing a system decide what applications would be run. In our institution, every student has a laptop, so we don’t have to run the students’ personal data on our system. We just have to allocate them 5 to 10 MB mailbox space depending on our learning management system,” he adds. Amity University, operating at an enterprise level, grew by more than 6,000 students in 2009. “Amity’s vision to provide the latest in terms of learning aids (e-labs, multimedia rich content, streaming video and demand) has increased the need for storage,” admits Dr J.S. Sodhi, certified security compliance specialist (CSCS) and assistant vice president of IT at Amity. Amity has a 24TB storage space mainly for folders, IP camera recording, virtual-class web-casting and archiving. “We have implemented server virtualisation at the Amity data centre through VMware with external shared storage,” says Sodhi.

DAS, NAS, SAN
According to Karapurkar, Direct Attached Storage (DAS) and Network Attached Storage (NAS) are commonly used options in the educational sector. DAS is the simplest system directly attached to a server or a workstation, without a storage network. It is connected to a computer or a server without a router or a switch. Even so, a DAS system can be shared among multiple systems, if it has interfaces that allow access. It can keep pace with high-storage demands and enable capacity extension according to needs. However, it means that an institute ends up creating islands of information, or information silos, across IT infrastructure. Every DAS device acts independently, so a centralised view is not easy.

NAS can be connected to a computer network. It allows file-level sharing in a network between systems. When this is deployed, the NAS setup takes responsibility of file serving from the server. NAS offers benefits if you deploy it with the right technologies. It allows multiple computer storing and effectively acts as a single repository. It can increase speed of access and storage of data compared to a server that runs other applications, thus slowing down processes. With the right RAID technology, one can build in redundancy of disks as well. However, as the number of users (accessing the NAS devices) rises, performance dips. An important factor is to check with vendors regarding the design.

Some users may need to go for a Storage Area Network (SAN)—network of storage devices that may be located at different rooms, but are visible at a single place as if locally attached to a server. There is the possibility of having disk arrays and tape libraries in SAN setups that enables storage sharing. A SAN is a better bet if disaster recovery is critical, for instance in places involved in highend research.

Cost is the major factor that drives system deployment. Karapurkar says, “We suggest either using DAS or NAS servers. SAN system is deployed for high-end research institutions that cannot afford to lose its data.” While a 1TB advanced NAS system may cost around Rs 200,000, a 1TB SAN box of similar or lower technology will cost around Rs 800,000. The difference is not limited to cost alone, but to complexity and manageability issues. SAN systems are a complex network of multiple boxes and require dedicated personnel to manage. Vivekanand Venugopal, the vice president and general manager of Hitachi Data Systems, says, “Hitachi Data Systems has addressed the complex nature of the SAN system and come up with simple SAN boxes that can be managed easily. We also provide adequate training and guidance to customers. Though the cost is high, institutions should not only look at the cost per TB, but also see the value that can be derived out of the setup.” Solution providers can also design hybrid systems using SAN and NAS capabilities that allow institutions to reduce cost and helps them to sort out the manageability issue.

A version of SAN called IP-SAN is a possible option, as it runs on regular Ethernet network without requiring fibre or a copper backbone. A key aspect is data accessibility—the speed and time that an institute is willing to spend to get a file off a storage server or disk. If speed is of essence, it is best to opt for an architecture that enables it—at a price. Higher data access speeds, clustering and load balancing are areas that need to be investigated before making a choice. Scalability is also important. A sense of how storage needs are expected to grow is cardinal to figure out the right kind of storage.

When it comes to bulk storage technology, tape and hard-disk drives rule the roost. Solid state storage devices are rare even at the enterprise level, as they are expensive. As for backup, mechanical ones are regular, while to automate the process, a software may be deployed.

Data Redundancy
Most data centres are comfortable with RAID5 that ensures data security even if a drive fails. RAID6, a more advanced technology, ensures that data remain intact even if two drives in a single array fail. “Amity implemented an online disk duplication back-up system, with two similar sizes of storage, ensuring complete redundancy. As size (of critical data) and demand increases, we will implement the High Availability Mode. In case of problems in one box, the second one will take up the session. In a normal state, both will distribute load. We also intend to implement remote disaster storage management,” Sodhi says.

Cost Worries
Cost is the major factor that decides storage setup selection. First of all, an approximate budget is needed to set things in motion. Calculation may be specific setup and the applications it will be running. Also, a course checklist and its storage requirement have to be made. Some of the questions that institutes need to ask themselves before zeroing in on a specific system: Is data storage only required for administration? Will students receive mail storage space, and if yes, how much? And that they may be used for classroom exchanges and faculty activities.

Once all this is done, a decision can be taken. Scout for the best deals. Consult partners for the deployment—as partners can double up as a consultant and help in deployment of a package without the institute having to hunt for every part. Karapurakar says, “Education as a sector enjoys the maximum discounting. Right from the vendor to the business partner, all are willing to extend their best deals to an institute.”