While FCoE technology has been standardized for quite some time now, most FCoE deployments have been upto the access layer of the network. Multi-hop FCoE deployments are gaining traction increasingly. Many a times, I get asked to share the production deployment designs and the real-world benefits of Multi-Hop FCoE infrastructure. So, in this series of blogs, I plan to share the same. In this blog, the spotlight is on a division of the world’s largest aerospace company, Boeing Defense, Space, and Security (BDS).
BDS provides end-to-end services for large-scale systems and supports a diverse range of customers, including the U.S. Army, the Securities and Exchange Commission (SEC), and U.S. Department of Defense (DoD). When the global recession hit the aerospace industry in 2010, BDS came under extreme pressure to cut costs. Dual network topologies, both FC and IP, were adding complexity to the network. BDS needed to reduce the TCO and at the same time increase the network agility, improve scalability and maintain highest availability possible.
As a result, the company decided to adopt FCoE to consolidate its IP and SAN data traffic on a single network. Since 2010, BDS has extended its use of FCoE and is now 100 percent Multi-hop FCoE. BDS deployed End-to-End FCoE architecture with Nexus 5000 at the access layer, the Director-class Nexus 7000 at the Core, connected to the FCoE Storage Arrays.
For BDS, the shift to the new Cisco Unified Fabric infrastructure and leveraging FCoE has delivered unparalleled value to the organization. Read More »
At some point, every data center has to migrate a complete server identity between two servers. This could be driven by maintenance needs, server upgrades or DR/HA and SLA requirements. For DR/HA, Business Continuity requirements mandate that this be done as quickly as possible, which means automation is critical to drive time to productivity restoration. True automation equals fewer steps and faster implementation, with the smallest possibility for error. This is more than just setting up a similar server with the same BIOS and firmware and then doing a lot of manual work requiring multiple administrative domains – compute, network and storage. I’m talking about QoS, vNIC / vHBA setting, storage access and ownership transfers, etc., the whole enchilada, all delivered in an automated repeatable processby design not by accident.
Not surprisingly, Cisco UCS with UCS Manager does the job -- fast, complete and in full. What may be surprising is that Cisco UCS Manager enables you to do this transfer not just from blade server to blade server, you can also do it from blade server to rack server. UCS Manager comes with and is embedded on the UCS Fabric Interconnects. I want to emphasize that there is no additional charge for UCS Manager, which is an important consideratin when you look at other companies’ multiple toolsets, agents and databases, most of which carry an additional cost, and which are required to equal UCS Manager functionality. UCS Manager architecture does not require a separate management server which other designs typically require.
The very best part of the entire activity is that the full migration of the server identity (enabled by Cisco SingleConnect technology) takes just 6 initial steps with UCS Manager; the rest is all about how we deliver on the promise of automation. UCS Manager lets you use and specify 127+ different server identity parameters including:
48 BIOS Settings
Host BIOS Firmware
Hdwr NIC Teaming (fabric failover)
FC Adapter & Storage Controller Firmware
BIOS & Disk Scrub Actions
Dynamic vNICs (VM FEX)
NIC and HBA Adapter Settings
vHBA WWPN & WWNN Assignment
HBA FC SAN Membership,
NIC Receive Rate & NIC MTU size
FC & iSCSI Boot Parameters
PCIe Bus Scan Order and PCIe Device Slot Placement
And Much Much More…….
The above all sounds good. Now we need to see ‘proof of delivery’. Below are the links to a white paper by Principled Technologies that are the real point of this blog – complete (not partial) migration of a server identity from a blade server to a rack server.
The real fun is to watch the accompanying video below. See for yourself how much time it takes in an apples to apples server identity migration from a blade to a rack server. Once you take a look at the video (the paper on the right has the full details of the testing), you will find taking a UCS Test Drive worthwhile.
The ability of Cisco UCS server to manage both blade and rack servers with a single tool is UCS Manager. Cisco took a unique approach to computing and focused on the common point of interaction, the fabric. Servers don’t operate in isolation. They are part of a total environment that at the minimum encompasses servers, networking, management and storage – a Fabric Based Infrastructure . Cisco’s comprehensive and efficient architecture is the key to why customers worldwide are rapidly adopting UCS.
For information on how UCS and UCS Manager integrate with a wide variety of our leading management partners follow this link UCS Manager Ecosystem Partners, and for interoperability with other major systems management tools please see the UCS Interoperability page.
Fast changing business conditions require agility, a difficult challenge in your distributed on-premises, big data and cloud environments. Data virtualization makes it easy for you to access your data, no matter where it resides.
Cisco’s integrated data platform optimizes query, compute and network infrastructure,so you access and query all types of data across the network as if it is in a single place.
You get the benefits of greater business insight and the flexibility you need in IT, with significant cost savings. You can then adapt to change more quickly and make better decisions in real time, without physically moving your data.
Data virtualization makes it possible to:
Empower your people with instant access to all the data they want, the way they want it
Respond faster to your changing analytics and business intelligence needs
While there is not yet an industry standard benchmark for measuring the performance of Hadoop systems (yes, there is work in progress -- WBDB, BigDataTop100 etc), workloads like TeraSort have become a popular choice to benchmark and stress test Hadoop clusters.
TeraSort is very simple, consists of three map/reduce programs (i) TeraGen -- generates the dataset (ii) TeraSort -- samples and sort the dataset (iii) TeraValidate -- validates the output. With multiple vendors now publishing TeraSort results, organizations can make reasonable performance comparisons while evaluating Hadoop clusters.
We conducted a series of TeraSort tests on our popular Cisco UCS Common Platform Architecture (CPA) for Big Data rack with 16 Cisco UCS C240 M3 Rack Servers equipped with two Intel Xeon E5-2665 processors, running Apache Hadoop distribution, see figure below, demonstrating industry leading performance and scalability over a range of data set sizes from 100GB to 50TB. For example, out of the box, our 10TB result is 40 percent faster than HP’s published result on 18 HP ProLiant DL380 Servers equipped with two Intel Xeon E5-2667 processors.
While Hadoop offers many advantages for organizations, the Cisco story isn’t complete without including collaborations with our ecosystem partners that enables us to offer complete solution stacks. We support leading Hadoop distributions including Cloudera, HortonWorks, Intel, MapR, and Pivotal on our Cisco UCS Common Platform Architecture (CPA) for Big Data. We just announced our Big Data Design Zonethat offers Cisco Validated Designs (CVD) -- pretested and validated architectures that accelerate the time to value for customers while reducing risks and deployment challenges.
Welcome back to the final episode of Engineers Unplugged, Season 3! It’s been quite a ride. This week, we take another viewpoint on the hot topic of software defined storage with Mike Slisinger (@slisinger) and Vaughn Stewart (@vstewed). Starting from the application owner’s perspective, this is a great 101 on the choices made on the road to the data center of the future. Let’s listen in:
Better stick to storage, not unicorns! Art by Mike Slisinger and Vaughn Stewart.
Welcome to Engineers Unplugged, where technologists talk to each other the way they know best, with a whiteboard. The rules are simple:
Episodes will publish weekly (or as close to it as we can manage)
Submit ideas for episodes or volunteer to appear by Tweeting to @CommsNinja
Practice drawing unicorns
How far up the unicorn scale is your data center in regard to software defined storage? Post a comment below!
Thanks for your viewership and support of Engineers Unplugged. We’ll be on site at VMworld Barcelona, camera and whiteboard markers in hand. If you’ve got show ideas or questions, tweet me @CommsNinja.