How Does the ECASD Network Work?

Slide 1 How Does the ECASD Network Work? Jim Blodgett, Network Engineer

Slide 2 Network Overview The ECASD Network has 3500 computers, 350 switches, 100 servers and 13000 users spread over 22 different buildings connected via lease fiber from Charter and our own fiber. We connect to the Internet through a WiscNet POP at UWEC on 2 pairs of fiber that take redundant paths around the center of the city. The network also supports all internal phone traffic and aggregates all incoming and outgoing calls over four T1 Primary Rate Interface (PRI) connections.

Slide 3 What Happens when I log in? ECASD uses Active Directory, a Network Operating System (NOS) from Microsoft, to manage and provide an interface to all of the resources connected to the network. Active Directory is also used to apply security restrictions and interface enhancements to users and computers in the form of policies. These policies ensure that the computers provide users with a uniform interface including Start Menu, Desktop, Storage and Application features and behaviors. These policies also help prevent users from making unsupported changes to the computers or network. All activities on the ECASD network, including keystrokes, are tracked and archived.

Slide 4 Where is my data? Home Folder School Folders Shared Folders Storage Area Network When a user logs on to a computer, all of the data and resources associated with that user is presented over the network by servers connected to a Storage Area Network (SAN). The Servers and the SAN are located in the Network Operations Center in the basement of the Central Office Building. This allows for very efficient management of the hardware and data while providing students and staff with a very stable and robust computing environment.

Slide 5 Why do we use a SAN? Normal computer One to several hard drives Independent Drives Inefficient Slow Limited space (<1TB) Limited reliability No redundancy Access limited to single computer Inexpensive When the hard drive fails the computer is down and the data is hard to recover Storage Area Network (SAN) Dozens of hard drives Drives work as a Redundant Array of Independent Drives Efficient: only deploy what is needed Huge amount of space (100 stb) Very fast: data can be read and written to all drives at once Data available to all network users Very reliable: can lose mulitple drives with losing data and servers remain available Data is consolidate making backup very efficient Expensive ECASD utilizes a Storage Area Network to manage data efficiently, providing a high level of performance and reliability. A normal computer can have several hard drives, though they usually only have one. These drives will be relatively slow (7200rpm on a SATA or IDE connection) with limited capacity and reliability. Filling a drive beyond 80% comes with a significant performance penalty. Without network storage, a user would need to use the same computer every time and backups would be inordinately complex and unreliable. In a SAN, all of the drives are aggregated into one big pool. Volumes are carved out of this pool with very little reserve space, and assigned to servers that provide the network interface. Data stored here is accessible from any computer on the network. Instead of slow, unreliable drives, the SAN uses very fast (15000rpm on a 4Gb/sec SCSI connection), very reliable hot-swappable drives that are setup in an array to provide high performance, high availability and redundancy. Because the data is consolidated, the SAN also allows for very efficient backups and recovery.

Slide 6 The ECASD SAN The center rack contains all of the disk arrays, the controllers, the management appliance and the switches that create the network fabric to tie everything together. The left rack contains the network tape library for backups and the home directory servers. The right rack contains the shared storage servers, the Exchange server and the database server. In the bottom of each rack are 2 generator compatible Uninterruptable Power Supplies (UPS) that provide electricity during the period between an electrical outage and the generator engaging.

Slide 7 Redundant Array of Independent Disks RAID takes two to many disks and ties them together to make one big volume RAID 0: data is striped across the disks. Very Fast no resiliency No lost capacity When one disk fails, the volume fails RAID 1(+0): data is mirrored across the disks Slower Resilient 50% capacity is lost RAID 5: Data is striped across the disks with a parity bit Need at least 3 disks Very Fast Very resilient Capacity is (Number of disks 1)*size of smallest disk What is RAID? All hard disks fail; they have moving parts and they wear out. When a single disk fails it is very hard to recover the data and the computer is not usable during that time. A Redundant Array of Independent Disks (RAID) provides varying levels of resiliency, efficiency and performance. In general, the more disks in the array, the higher the resiliency, efficiency and performance, with the exception of RAID 0, which provides very high efficiency and performance but no resiliency. In RAID 0, the disks are tied together in one volume and the data is striped across all disks, so the loss of any disk causes the volume to fail. In RAID 1, the data is written to the first disk and then mirrored to the second disk. This actually slows down the computer, but provides a high degree of resiliency, though a failed disk needs to be replaced quickly. It is also only 50% efficient with regards to space. In RAID 5, the disks are tied together and striped, like in RAID 0, but a parity bit is also written, so the volume can be reconstructed without going offline, in the event of a disk failure. This is very fast and highly resilient, but costs one disk out of every array for parity, so the more disks in the array the more efficient it is. The loss of 2 disks will cause the array to fail. Our SAN uses a proprietary impelmentation of RAID 5 that provides much more resiliency than standard RAID 5.

Slide 8 What happens when I surf the web? ECASD must meet the requirements of the Child Internet Protection Act (CIPA) and the Deleting Online Predators Act (DOPA) to receive Erate funding. That means we must filter Internet content and provide students with information about how to protect themselves when they are on line.

Slide 9 How does my email get filtered? Email is the vector of choice for distributers of unsolicited advertising (SPAM), malware and viruses. We receive roughly 3 million emails a day, 96% of which are from known spammers. Of the email that is not from a known spammer, half is classified as spam or malicious, so we get roughly 60, 000 legitimate emails a day.

Slide 10 Questions? Please refer any questions that your teacher can not answer to: Jim Blodgett jblodgett@ecasd.k12.wi.us