Cisco Blogs


Cisco Blog > Data Center and Cloud

UCS M-series Modular Servers – Because wastage just plain hurts!

Last week we announced the UCS M-series Modular Servers. The launch represented culmination of an exciting journey for us that started two years ago.

In mid 2012 just as UCS B-series blade servers were taking off in a big way, we noticed a group of our customers using our core technology very differently than customers in our primary market, enterprise IT. In our primary market customers loved UCS’s stateless computing model, virtualization benefits and the converged offerings with our partners EMC and NetApp. In this other category, customers did not consider those same benefits nearly as important. However UCS Manager’s powerful policy engine got them really excited. UCS Manager gave them a programmatic interface to manage thousands of nodes across dozens of sites globally.

Curious, I started to visit some of these customers. During one such visit, I was walking thru the aisles of their data center and I noticed something I had not ever seen at any of our enterprise IT customers data center. This customer had all UCS chassis single homed to a single Fabric Interconnect, I stopped in my tracks -- reallyIsn’t that kind of dangerous? What happens if there’s a failure? Or you have to upgrade? The customer explained to me how a combination of their application architecture and their application instance placement strategy made sure that outages at the rack level could be handled without service disruption. Wow! so we had engineered all kinds of resiliency, dual ported adapters, dual IOMs, dual chassis controllers, clustered Fabric Interconnects … lots and lots of hard engineering work to make our product robust and resilient, and this customer had thrown it all away with one toss… that really hurt. :( Read More »

Tags: , , , , ,

#EngineersUnplugged S6|Ep8: #UCSGrandSlam Edition!

September 10, 2014 at 9:34 am PST

In this week’s episode of Engineers Unplugged, Cisco’s CTO, Padmasree Warrior (@padmasree) and Satinder Sethi (VP, UCS Product Management and Data Center Solutions) whiteboard the UCS Grand Slam announcement, and what it means for customers and for the modern data center. Don’t miss this one!

It wouldn’t be Engineers Unplugged without a unicorn challenge, and Padma and Satinder delivered!

10669318_10152774542251349_1106876536311792259_o

10669318_10152774542251349_1106876536311792259_o

This is Engineers Unplugged, where technologists talk to each other the way they know best, with a whiteboard. The rules are simple:

  1. Episodes will publish weekly (or as close to it as we can manage)
  2. Subscribe to the podcast here: engineersunplugged.com
  3. Follow the #engineersunplugged conversation on Twitter
  4. Submit ideas for episodes or volunteer to appear by Tweeting to @CommsNinja
  5. Practice drawing unicorns

Join the behind the scenes by liking Engineers Unplugged on Facebook.

Tags: , , , , , , ,

Data Vault and Data Virtualization: Double Agility

Rick van der Lans is data virtualization’s leading independent analyst.  So when he writes a new white paper, any enterprise that is struggling to connect all their data (which is pretty much every enterprise), would be wise to check it out.

Rick’s latest is Data Vault and Data Virtualization: Double Agility. In a nutshell, the paper addresses how enterprises can craftily combine the Data Vault approach to modeling enterprise data warehouses with the data virtualization approach for connecting and delivering data.  The result is what Rick calls double agility as each approach accelerates time to solution in complex data environments.

Data Vault Pros and Cons

Adding new data sources such as big data and cloud to an existing data warehouses is difficult. The Data Vault approach provides the extensibility required.  This is the first agility.

Unfortunately, from a query and reporting point of view developing reports straight for a Data Vault‐based data warehouse results in complex SQL statements that almost always lead to bad reporting performance. The reason is Data Vault models distribute data over a large number of tables.

Losing Agility Due to Data Mart Proliferation

To solve the performance problems with Data Vault, many enterprises have built physical data marts that reorganize the data for faster queries.

Unfortunately valuable time must be spent on designing, optimizing, loading, and managing all these data marts.   And any new extensions to the enterprise data warehouse must be re-implemented across the impacted marts.

Data Virtualization Returns the Agility

To avoid the data mart workload, yet retain agile warehouse extensibility, Rick has worked with Netherlands based system integrator Centennium and Cisco to provide a better, double agility, alternative.

In this new solution, Cisco Data Virtualization, together with a Centennium-defined data modeling technique called SuperNova, replaces all the physical data marts.  So, no valuable time has to be spent on designing, optimizing, loading, managing and updating these derived data marts. Data warehouse extensibility is retained, but because the reporting is based on virtual, rather than physical models, they are very easy to create and maintain.

Meet Rick van der Lans at Data Virtualization Day

To learn more about this innovative solution as well as data virtualization in general, come to Data Virtualization Day 2014 in New York City on October 1.  Rick, along with the also sharp Barry Devlin, will join me on stage for the Analyst Roundtable. I hope to see you there.

 

Learn More

To learn more about Cisco Data Virtualization, check out our page

Join the Conversation

Follow us @CiscoDataVirt #DVDNYC

Tags: , , , , ,

Cisco UCS M4 Compute Platforms: Performance That Matters

Part of last week’s UCS Grand Slam launch event last week in NYC was the announcement of three new compute platforms – Cisco UCS C220 M4, C240 M4, and B200 M4.  Today, Intel announced the new Intel Xeon E5 v3 family of CPUs that will power these new UCS platforms.  Hopefully, this week, now that the confetti has settled from our brand new groundbreaking products like the M-Series with System Link Technology, UCS Mini with the new UCS 6324 Fabric Interconnect, and the capacity-optimized C3160 Rack Server, we can highlight some updates to our core compute platforms.

Its easy to get caught up in the new platforms, they are the new vehicles to bring the benefits of UCS to new markets and at a scale that was previously impractical.  But it’s important to remember that the UCS two-socket blade and rack servers were the original foundational platforms to bring the benefits of UCS to the datacenter.  In fact, the predecessor’s to these products propelled UCS to some amazing accomplishments.

Accomplishment

So, let’s pull back the covers a bit more on these very capable foundational compute platforms that make up many of the building blocks for the enterprise datacenter.

First, when we began to design the latest version of each product, we set out to follow a few simple rules.  Principle of these was to understand what makes them so popular and enhance those elements.’  For example, the B200 M3 is the best-selling server blade in the product line mainly due to its amazing versatility and uncompromised feature-set, all while maintaining a half-width form factor.  Well, the UCS B200 M4 Blade Server is more of the same and then some.  It still delivers uncompromised features like the highest speed, core count and TDP CPUs, maximum memory with 24 DIMM sockets and industry-leading I/O with support for both 2nd and 3rd generation Cisco Virtual Interface Cards (VIC) at up to 80Gb/s of bandwidth per blade.  And, all of that can be done simultaneously, no compromise.

UCS B200 M4 Blade Server

UCS B200 M4 Blade Server

That was easy enough to deliver, but in order to enhance this platform, we looked at the use cases that it served and found the flexibility was the next pivot point for the B200 M4.  The addition of Cisco Flexible Storage to the B200 M4 means that now, customers can truly scale the storage subsystem to match their needs.  Today many UCS customers take full advantage of the true stateless computing and do not use local storage on the blade.  For those use cases, it may be appropriate to have no local disk, no local RAID controller and even no local disk bays.  Why pay for, power and cool what you do not use?  For still other applications, not only is local storage critical, but high performance SSD with an equally-high performing RAID controller with flash-backed write cache is needed.   This is where the Flexible storage subsystem shines.  Either extreme, and the points in between are covers and no compromises made elsewhere.

Another tenet of our design philosophy was to focus innovation where the use cases could take advantage of it.  Take the UCS C240 M4 Rack Server for example, its M3 predecessor has found a home in many enterprise workloads, but it main differentiation is its optimization around local storage and I/O.  To that end, the C240 M4 has enhanced storage flexibility features that include a modular RAID Controller with optional Flash-backed write cache, options for up to 24 SFF or 12 LFF front-accessible hot plug HDD / SSD, two additional SFF internal boot drives, and even support for two 2.5” PCIe flash devices in the front drive bays.

UCS C220 M4 Rack Server

UCS C240 M4 Rack Server

The I/0 capabilities are also significantly improved with up to six PCIe slots that can house up to two NVidia Kepler GPUs.  We also added an mLOM. This slot that is optimized for VIC or 3rd party network cards add to the embedded GbE NIC ports without using one of the six PCIe slots.  The C240 M4 is the ideal platform for I/O and storage intensive enterprise bare metal and virtualized workloads.

The new UCS C220 M4 Rack Server shares a similar compute engine with the ability to also support up to 18 cores per socket and 1.5TB of memory, but it offers an optimization around density without compromising the enterprise performance and feature-set.  So you get the same enterprise class power and versatility in a compact footprint.

UCS C220 M4 Rack Server

UCS C220 M4 Rack Server

These new platforms have already begun leaving their mark with four new world record benchmarks that are targeted toward showing real benefit when deploying workloads that bring tangible value to our customers on platforms that customers demand.  This is how UCS Servers deliver performance that matters.

Look for more information on UCS innovation in this space in the future.

Scaling Application Security with ITD

Ready to scale your enterprise beyond limits?  How about slashing a whole layer of datacenter infrastructure, saving piles of cash in the process?  Or perhaps you’re interested in simplifying your enterprise while adding features, or trying to speed things up without spending money.  Sound too good to be true?  Well, thanks to a new technology from Cisco, you can have your cake and eat it, too.

Cisco Intelligent Traffic Director (ITD) is poised to disrupt data center load balancing. Combined with best-in-class products, such as Imperva SecureSphere, organizations can deploy and manage massively scalable applications securely with unprecedented ease and cost effectiveness.

What is ITD?

Cisco recently released a new feature, Intelligent Traffic Director (ITD) for the Nexus 7k switches that promises to be a disrupting force in the world of load balancing.  There has been an exponential growth in data traffic in the recent years leading to a growth in the deployment of network service appliances in enterprise, datacenter and cloud environments. To address the corresponding business needs, network switch and router architecture has evolved to support multi-terabit capacity. However, service appliance capacity remained limited to few gigabits, an order of magnitude far below switch capacity.

Cisco Intelligent Traffic Director (ITD) is an innovative solution that tries to bridge performance gap between the switch and service appliance(s). It allows customers to deploy service appliance(s) from any vendor with no network or topology changes. With a few simple configuration steps on a Cisco Nexus 7000 / 7700 series switch, customers can create a service appliance cluster and deploy multiple appliance(s) to scale service capacity with ease. The servers or appliance(s) do not have to be directly connected to the Nexus switch.

Application Security
Gartner published a paper called Web Application Firewalls are Worth the Investment for Enterprises in Feb, 2014 that makes the case that “Firewalls and intrusion prevention systems don’t provide sufficient protections for most public-facing websites or internal business-critical and custom Web applications.” Gartner advises enterprises to use a Web Application Firewall (WAF) to protect critical external and internal applications from attacks and threats.

Like other service appliances, a WAF appliance benefits from ITD’s ability to manage large scale traffic loads. Imperva SecureSphere WAF works with ITD, and the combination provides highly scalable application security.

I mention SecureSphere because Imperva was positioned as the only Leader in the Gartner 2014 Magic Quadrant for Web Application Firewalls. Some key capabilities of the SecureSphere WAF are:

  • Block attacks with laser precision
    Accuracy is critical with application security. If you have false positives, you block customers; if you have false negatives, you let the bad guys in.
  • World-renowned application security research
    Security is constantly evolving. To get ahead and stay ahead in the continuous fight against threats, Imperva has a dedicated security research team, the Application Defense Center (ADC), which provides regular signature and policy updates, and up-to-date threat intelligence for Imperva SecureSphere.
  • Shut down malicious sources and bots
    Imperva’s ThreatRadar Reputation Services help detect bad actors using IP reputation feeds of known malicious sources, anonymizing services, phishing URLs, TOR (“The Onion Router”), as well as IP geolocation data.
  • Stop application DDOS and business logic attacks
    Business logic attacks include things like posting comment spam in forums and message boards, scraping web content, and disabling access to your website. All of this can reduce competitive edge, frustrate customers, and damage reputation.
  • Instantly patch website vulnerabilities
    It takes organizations an average of 6 months to patch an application vulnerability once it’s discovered. SecureSphere integrates with vulnerability scanners to virtually patch applications. This allows businesses to stay protected, and fix the vulnerability on their own timeline, thus reducing the window of exposure and the associated costs.
  • Gain forensics insights with customizable reports
    Graphical reports enable organizations to quickly analyze security threats and meet compliance requirements.
  • Speed up deployment without risk
    SecureSphere protects applications without impacting performance and without requiring extensive network changes. It offers flexible inline, non-inline, and proxy deployment options that meet organizations’ diverse requirements. SecureSphere’s Fail-Open capabilities combined with unique, transparent bridge mode saves time and labor with drop-in deployment that requires no changes to existing applications or network devices, and delivers multi-Gigabit throughput while maintaining sub-millisecond latency.

Scaling Application Security

Using ITD in VIP Mode to load balance provides a fast and economical way for organizations to provide highly scalable and available infrastructure.  By leveraging ITD, an enterprise can deploy a single IP address (the VIP), which is then load balanced across many SecureSphere WAFs, with each one protecting the back-end webservers. This is done right from the 7K – There’s no need for an external load balancer in the middle.

Why is this better than other Load Balancers?

By combining Cisco ITD and SecureSphere’s advanced capabilities to monitor and secure HTTP traffic, several key advantages are apparent:

  • Eliminates the need for external load balancers, freeing up large amounts of budget and resources
  • You get the advantages of a proxy-type load balancer (1 single VIP represents many webservers), but still get ‘fail-open’ bridges on WAFs
  • ITD proxies traffic without interfering with the TCP Source IP Address , allowing SecureSphere to leverage the source IP, User and Session details for blocking and alerting.
  • To work with SecureSphere, ITD requires no modification to HTTP Headers (e.g., X-Forwarded-For), which can break applications and slow down traffic

What does this mean for the future of high performance WAF deployments?

By teaming up the Cisco Nexus 7K with SecureSphere WAFs, organizations can cost effectively deploy scalable, high-availability  WAF farms to handle large amounts of traffic to webservers.  As the web traffic increases, WAFs can be seamlessly added to the pool to scale up with the enterprise. Since every port on the 7K can be used as a load balancer this provides the potential to scale up to multi-terabits of throughput to a SecureSphere WAF cluster.

In conclusion, ITD and SecureSphere provides simple, cheap, fast, scalable, and reliable security infrastructure. Sort of like having your cake, with icing, and cherries on top – and eating it, too.

Feedback or Query: For feedback, query or EFT/PoC/demo please email: ask-itd@external.cisco.com

ITD White paper:  At a glance

Configuration Guide: Config guide

 

Tags: , , , , , , ,