HP Network Node Manager 9.10i - A Quick Analysis

HP Network Node Manager 9.10i - A Quick Analysis
Abstract:
For as long as there was communications over wires, Network Management existed. With the advent of SNMP, management could become sophisticated enough to allow for proactive management of intelligent devices and the health of various resources they managed. HP's Network Node Manager, considered "the gold standard" in network management, continues to upgrade their platform with newer features.

Key Information:
Real information on Network Node Manager is available in various datasheets and specification pages. What separates HP's Network Node Manager from the rest of the competition includes:

• Multi-Platform Support
  Solaris, Linux, Windows, and HPUX.
  Solaris support provided for Managed-Services grade, Linux for Enterprise grade, Windows for Small-Medium Business grade network management needs.
• Robust Performance
  Single station capable of 25,000 devices, up to 1,000,000 interfaces.
• Robust Virtualization
  HP Network Node Manager supports Zones under Solaris - providing Solaris based Managed Service Providers with secure multi-tenancy at 0% systems overhead penalty.
• External Authentication
  Users and Roles can be mapped to an LDAP directory
• Excellent Device Certification Documentation
  Clear on-line documentation regarding which device and features are supported.
• Database Support
  Embedded Postgres and third-party Oracle are both supported.
• Multi-Tenancy Interface
  Network Node Manager old interface offered this, the new one offers this, as well.
• Encryption and SNMPv3
  The old Network Node Manager did not offer SNMPv3 encryption, without a proxy from SNMP Research. The data sheets do not clearly indicate SNMPv3 encryption support under multi-tenancy and security.
  

Analysis:
Everyone who wants to be a managed services provider plays in the Linux space, but only development houses truly capable of supporting large telco environments play in the UNIX space. Solaris is the dominate market leader, with more installations than IBM AIX and HP-UX combined, but HP-UX is still supported (for the time being - their Itanium based HP-UX platform is at-risk.)

Scalability of 25K devices with 1M interfaces is a terrific improvement, over the old days of HP Network Node Manager. The older NNM platforms seemed to become sluggish between 1,000 and 2,000 devices. Modern day telco managed services engagements regularly deal with multiple core devices have 1,000,000 interfaces per core device (not to mention end-points) - which this writer was able to do a half-decade ago with EMC Ionix on an old 8 socket UltraSPARC IV platform (as well as other customers.) Running a separate NNM instance for every internal core device is completely unrealistic. Most people do not consider the vertical scalability needs of network management tools - single customers with large interface counts MUST support large socket counts on UNIX systems like Solaris and HP-UX to meet the business requirements.

With 0-overhead virtualization technologies like Solaris Zones, Node Manager is able to play "in the big leagues" with managed service providers, to provide simplistic multi-tenant solutions where overlapping ip addresses must be managed. Each Solaris Zone can route to a different customer, with identical IP Addresses - with no cost in system resources, since Solaris Zones all share the same kernel resources.

Authentication of users to an external LDAP directory provides for Single Sign-On capabilities and resolves a variety of issues with auditing requirements, password changing issues, locking out users who are no longer with the service provider, etc.

It is critical for a managed services organization to understand whether a device can be monitored and managed. Clear documentation, which is available for the world to see, provides that kind of feedback to the community, to level-set expectations with telco providers to communicate to joint customers.

Embedded Postgres and third-party Oracle are both supported, the former for service providers to avoid Oracle service provider fees, the latter for enterprises with few DBA  resources to provide singular database support across an entire enterprise.

Multi-Tenant requirements are tremendously diverse. The new web interface offers configuration, like the old X-Windows interface, but the data-sheets are strangely silent about SNMPv3 support - an absolute minimum for dealing with multi-tenancy, so customers know their data is secure during the traversal of shared networks. Any security minded individual be concerned about support or robustness of support of SNMPv3 by this product.

Additional Information
See the SNMP Resources Tab on the Network Management Blog for more information.

Oracle Solaris 11: Session 4: Core to the Oracle Systems Strategy

Oracle Solaris 11: Session 4: Core to the Oracle Systems Strategy

Online Forum
Oracle Solaris 11:
What's New Since the Launch
April 2012

Abstract:
Oracle released a series of 4 sessions on Oracle Solaris 11: What's New Since the Launch (see Session 1, Session 2 and Session 3.) Oracle Executive Vice President of Systems, John Fowler, elaborates on how Solaris is the core to the Oracle systems strategy. Notable was the fact that there are more Solaris instances in the marketplace than AIX and HPUX combined and engineering is pushing SPARC and Solaris into the future with ever sophisticated systems requirements.



Executive Vice President of Systems
John Fowler

Oracle is investing in 3 ways:
1) servers and storage
2) enterprise core technologies
(testing and problem resolution, incorporate software)
3) oracle engineered systems
(server, storage, networking, and Solaris integration)


Engineered Systems: Exadata, ExaLogic, SPARC SuperCluster
Acceleration of: Database, WebLogic, Java


Oracle SPARC T4 - Driving Outstanding Performance


Solaris unlocks key features to increase performance


Oracle Storage - Everything runs faster


Oracle Solaris - More installations than AIX and HPUX combined


Oracle Solaris - #1 UNIX for the Next Decade of Hardware


Top business application are under Solaris 11


One Engineering Team: Hardware, OS, Middleware, Applications

Processors for: AIX, HP-UX, and Solaris

Abstract:
With the delay and loss of single-vendor advocated central processor units, operating systems centered on that silicon are considered not necessary. Application vendors dependent upon a single operating system dependent upon a single underlying architecture are even at higher risk. This article discusses the top 3 UNIX vendors with the impact of their silicon trajectory.


[Rumored IBM Power 7+ Multi-Chip-Modules, courtesy softpedia]
IBM POWER 7+ Now 7 Months Late: AIX in the Crosshairs

It was reported last month that IBM POWER 7+ was delayed about 6 months. As of this date, POWER 7+ is now 7 months late. A little insight from the IT Jungle from end of January:

This year, IBM is also supposed to add support for Power7+ machinery to the SDMC, which presumably implies that Power7+ processors are coming out sometime this year.

Maybe 7 months for IBM POWER 7+ is enough?
Will IBM POWER 8 be mentioned 8 months after POWER 7+ was late?

Of course, the dominant operating system on single-supplier IBM POWER is IBM AIX. The delay single-supplier of POWER impacts single-supplier IBM AIX. Linux may also run under POWER, but Linux is not the reason POWER exists.

Applications locked-into IBM AIX on IBM POWER must wait. Applications compiled for IBM POWER under Linux must also wait. If an business application needs more power, one must wait on POWER.


[Intel Itanium, courtesy xbitlabs]
Intel Itanium End: End of HP-UX
The Web Logic Development Journal listed some HP statements during the Oracle-HP court case:

• HP did not want to reveal that the Itanium road map is "more an illusion than of technical significance."
  
• Its purpose was to "extend the Itanium roadmap... to create market perception of long term viability."
  
• "HP-UX is on a death march due to inevitable Itanium trajectory."
  
• That HP knew that customers are prone to abandon a server technology as soon as its end of life becomes "visible"
  
• HP's internal documents show that "the Itanium situation is one of our most closely guarded secrets."
  
• "The regions are unaware of the situation with Itanium and the impending end of life."
  
• The last Itanium chip, Kittson plus, is released in a throttled down version and then a full version to create "illusion" of longer roadmap.

The end-of-life for (single supplier) Intel Itanium is not a surprise, but the court statements are interesting.

In November of 2011, 3000newswire discussed Project Odyssey, HP's delivering HP-UX features to only Linux and Windows.

"Unfortunately project Odyssey will ultimately drive most companies to IBM's AIX. [HP-UX] features on Linux are desirable, but Odyssey won't get many customers to migrate to Linux. I think it is very interesting that only Linux and Windows are supported.

Clearly, HP-UX is on the ropes. The [court document revealed] death of Intel Itanium co-insides with HP pushing users off HP-UX and Itanium. With Oracle shutting down Itanium software development, the death of Windows on Itanium, death of Red Hat Linux on Itanium - the push off of Itanium's sole remaining HP-UX operating system is not unusual.

[Oracle SPARC T4 Processor]
Multi-Vendor SPARC: Solaris Diversity Thrives
Customer dependent upon IBM POWER or Intel Itanium are locked into those vendors and their single source operating systems (i.e. IBM AIX and HP-UX)  as discussed earlier. The death-march by customers dependent upon those operating systems is uneasy. As single vendor CPU suppliers silently delay, operating system vendors tie their customer's fortunes to those single suppliers. Operating systems tied to single CPU vendors are even more at risk.

In the SPARC community, things is vastly different. SPARC is a specification, multiple commercial vendors build SPARC processors and systems, anyone can choose to make their own SPARC chips without going to existing vendors, there is no legal risk for additional vendors to building SPARC processors, and multiple operating systems by multiple profitable vendors exist for processor support.

Commercial SPARC vendors include: Oracle, Fujitsu. Fujitsu has the most diverse lines, including: Throughput, Mainframe, and SuperComputer SPARC models. Commercial operating system support for SPARC includes: Solaris, Linux. Solaris CPU support include: Fujitsu SPARC, Oracle SPARC, Intel x64, AMD x64. Solaris family OS vendors include: Fujitsu, Oracle, Joyent, Nexenta. Solaris based Open-Source distributions include: SmartOS, OpenIndiana, Illumian. Solaris source code trees include: Closed Source Oracle Solaris, Open Source snapshots of OpenSolaris; active Open Source Illumos.

Oracle released a 4 part virtual seminar on the Solaris 11 road map (note: published comments on Session 1.) With new processors from Fujitsu and Oracle being released (seemingly yearly) and diversification of Solaris under all major commodity processors - it feels like the 1990's, with the launch of the Internet, all over again!

IBM POWER Roadmap... 7+ now late and only an almost 3 years projection for 8?

IBM POWER Roadmap... 7+ now late and only an almost 3 year projection for 8?

An image speaks a thousand words, late today with a blue box for a future, just before the Oracle SPARC T4 release?


Thank you for the IBM August 2011 POWER Roadmap tha the public marketplace has been begging for... did we miss the POWER7+ release??? A POWER 7 February 2010 launch would have POWER 7+ August 2011 launch (and today is August 31, so unless there is a launch in the next 23 hours, it looks late to me.)

Sketchy details on something possibly 3 years out??? No commitment beyond (almost) 3 years for POWER???

POWER has been quite interesting, for being cobbled together via multi-chip modules. Certainly a less risky approach, but a more expensive approach which does not offer flexibility of mass-production. Of course, there was never a delusion in the standard POWER family of trying to "make it big" - those days are long gone, with the former IBM partnership with Apple and Motorola.

It is nice to see embedded POWER in some video game machines and cell phones, but we have not seen a commodity desktop chip in a long time.

SPARC Open CPU Architecture Roadmap

With multiple vendors developing SPARC, where anyone can develop SPARC processors, without resriction - IT executives and government organizations may feel more comfortable with another type of roadmap.

It would be nice if IBM the confidence in POWER that Oracle has in SPARC, who released a 5 year roadmap, where Oracle has been providing continual public updates as SPARC benchmarks have been met.

The new SPARC T processors have been on-time for every generational launch for the past 5 years, with Solaris Update 10 already leaked and T4 processors about to be released.

With the decision to invest in a processor based upon a single piece of silicon, the ability to mass-produce at lower cost created low-cost options for SPARC in the past, at a time when few companies were trying to produce low-cost, embedded and commodity processors.

With SPARC, we have not seen a commodity desktop workstation for the education and scientific markets in awhile, either. After the Ultra 45 Workstation and migration of the SunRay UltraThin Clients from MicroSPARC IIe, that was the last of a worthy line of systems.

The market is hoping to see something from another OEM vendor, since one might suspect Oracle is not terribly interested in anything other than servers, and there always seems to be an OEM vendor releasing a SPARC compatible portable or deskop. The upcoming T4 is a terrific candidate.

Network Management Considerations

There are multiple vendors who design and have been encouraged to design SPARC, from the United States, to the U.K., to Africa, to Europe, to Russia, to China, and to Japan. The SPARC application market is best when there are multiple governments, educational facilities, companies, and startups who are competing to produce a better (or, at least, available) design.

Markets with competition and open designs have always been better for the industry as well as the customers. There is a reason why Solaris is trusted in the telecommunications arena and has nearly 2x the number of applications than combined AIX and HPUX.