Regaining control of your company's data

Interestingly enough, what makes resolving this dilemma of centralizing the distribution of data so perplexing is physics itself--specifically, the speed of light. Without nature's "speed limit" on data traversing the WAN, all servers could be completely centralized and fully secured without sacrificing performance for distant users. Indeed, Gartner believes "server consolidation projects that fail to account for network latency will fail to deliver acceptable performance." There are other challenges encountered in the WAN, of course, most notably the cost of bandwidth and the reliability of connections, which must also be taken into account.

Despite these challenges, regaining enterprise-wide control of corporate data requires consolidation and centralization; there is simply no other way to ensure data integrity, comply with regulations, and fully contain costs.

To address this business-critical need, the networking industry has responded with a variety of WAN optimization technologies. These improvements in WAN throughput are a necessary first step in controlling data, but they are only part of what is needed.

The complete centralization and consolidation of data and IT resources has three priorities:

1. Continue empowering knowledge workers with fast access to business-critical information, including e-mail, Web and application data,
2. While capturing and securing 100 percent of the data being created in remote locations, and
3. Ultimately eliminating the "IT infrastructure" required at branch offices.

Provide Fast Access to Business-critical Information--The distribution of business processes and the dissemination of business-critical information must continue even as the data and IT resources are centralized and consolidated. Virtually any WAN optimization technology will suffice for this requirement, including wide area file services (WAFS), caching, compression, protocol spoofing, etc. But full centralization of information assets requires much more.

Capture and secure 100 percent of corporate data--Industry analysts estimate that from 60-75 percent of corporate data is generated and "maintained" in locations other than the corporate headquarters. To complicate matters further, a significant and increasing percentage of this data now resides at the extreme edges of the enterprise network on individual desktops and at the home offices of mobile workers. Ensuring the privacy, integrity and retention of all this data is not only a best practice, these and other provisions may now be required by the coercive force of law.

For example, information about individuals must be secured from unauthorized access under laws, such as the Health Insurance Portability and Accountability Act (HIPAA) in the United States, along with similar laws in other countries and some states. Some of this legislation, such as the Sarbanes-Oxley Act in the United States, mandates that certain information be preserved in routine back-ups. Then there is the problem of "silent data loss" that can occur when a WAFS-based system encounters an unnoticed error condition (e.g. a WAN disruption, media error or protocol timeout) that corrupts the data.

The key to successfully centralizing data is the "authoritative" version of every file. With WAFS, the authoritative version can be kept securely (and backed up routinely) in one centralized location. Solutions taking this approach should incorporate a distributed file system that utilizes a lock manager, along with robust journaling, recovery and replication techniques, to protect remote users from stale data, write collisions, WAN outages and system failures. The solution should leverage multiple performance-enhancing techniques, as well, including one-point storage caching, data streaming, read-ahead, file synchronization, file-aware differencing, compression, data aggregation, and I/O clustering to achieve LAN-like file access in the remote offices. Of course, the approach should also be transparent to the users.

Minimize remote IT infrastructure--The ultimate goal of centralization is the elimination of "IT infrastructure" in remote locations. Obviously, there needs to be some equipment in remote office/branch office (ROBO), small office/home office (SOHO) and "Road Warrior" environments. But in the case of consolidation, less is more. Gartner, for example, envisions the "Branch Office Box" or BOB that consolidates all required functionality in a centrally-managed appliance. BOB must perform tasks that are both visible (e.g. e-mail, file/print services and various applications) and "invisible" (e.g. name services, address assignment/translation, etc.). To accommodate these tasks, Gartner believes BOB will need to provide a range of capabilities, including WAN optimization, caching, data compression, WAFS, content distribution and more.

The benefits of the BOB are quite profound when centralizing enterprise IT. Remote users continue to have seamless, high-performance access to both local services and corporate resources. And the organization eliminates the need to deploy and manage critical IT infrastructure and data at each branch office, which simplifies administration and regulatory compliance, and affords substantial cost savings.

The benefits of controlling data through consolidation

• Lower overall IT costs, owing to the reduction of remote IT infrastructure and enhanced utilization of existing assets

• Full centralization of data management, including significant reductions in the storage management burden and backup/recovery times

• Better support of enterprise-wide best practices, policies and regulations pertaining to the retention, storage and retrieval of critical company data

• Higher levels of service surrounding data availability, reliability and protection Getting started
The Yankee Group provides some useful advice for the enterprise in its report titled Consolidating Branch-Office Infrastructure Optimizes Information Management and Protection. Because demands on the network will increase as organizations consolidate data management and optimize the WAN to deliver services from a centralized data center, Yankee Group suggests implementing the solution in phases.

Initial phases should demonstrate an immediate ROI by targeting file, print and Web servers. The initial phases should also begin the centralization of backups and other data protection strategies, including those required for retention and archiving. Later phases can then tackle the centralization of major corporate applications, such as e-mail--one of the largest management headaches in any enterprise. According to Yankee Group, "Consolidating e-mail infrastructure can return big dividends and really show the value of the project."

As for the solution itself, Yankee Group emphasize the need for an offering that delivers value today, and will adapt and grow in lockstep with branch-office needs. Specific capabilities identified include scalability, performance, reliability and resiliency, security and seamless integration with existing infrastructure.

Gartner offers some other useful advice in its research note titled BOBs Help You Make the Most of Branch Office WANs

• Develop a clear cross-functional plan that includes server, application, storage and networking teams. • Identify which services are required at each branch office and which ones must be available during a WAN outage. • Test the plan's assumptions and verify its technical feasibility under real-world conditions.

Regaining full control over all corporate data enterprise-wide through consolidation and centralization is well worth the effort, according to the Yankee Group: "By delivering consistent, reliable IT services to every branch, the enterprise doesn't just optimize its WAN; it optimizes its business." Indeed, without centralization, attempts to achieve regulatory compliance for and gain full control over information assets may forever remain elusive--and expensive.

biography
Shirish Phatak, founder of Tacit Networks and CTO of WAFS Products at Packeteer, leads the design and development of the company's wide area distributed file system technology and its integration into the company's NAS caching solutions. He has also contributed significantly to the Intermezzo file system project, which is now a part of the stock Linux kernel. Prior to founding Tacit, Mr. Phatak was a Ph.D. researcher with the Dataman Mobile Computing Research Labs in the Department of Computer Science at Rutgers University. Mr. Phatak also worked for the Office of Corporate Liaison and Technology Transfer at Rutgers University, where he managed the processing and drafting of various patents and negotiation of licensing deals for various IT technologies owned by Rutgers. He holds an M.S. in Computer Science from the Indian Institute of Technology, Mumbai, India (formerly IIT, Bombay) and a B.E. in Computer Science from the S.G.S. Institute of Technology, Indore, India.